Introduction to the Oracle Stream ExplorerFast Data
ORAC L E W H I T E
Introduction to the Oracle Stream ExplorerFast Data
ORAC L E W H I T E
Introduction to the Oracle Stream ExplorerFast Data and Event Processing
ORAC L E W H I T E P AP E R | M ARCH 2 0 1 5
Introduction to the Oracle Stream Explorerand Event Processing
MARCH 2 0 1 5
Introduction to the Oracle Stream Explorerand Event Processing without Software Coding
Introduction to the Oracle Stream Explorerwithout Software Coding
Introduction to the Oracle Stream Explorerwithout Software Coding
Introduction to the Oracle Stream Explorer without Software Coding
1
Table of Contents
Introduction
Understanding Shapes, Streams, References and Explorations
Case Stu
Conclusion
Appendix A: Scripts and Samples used in the Case Study
Appendix B: Creating a Message
1 | INTRODUCTION TO THE
Table of Contents
Introduction
Understanding Shapes, Streams, References and Explorations
Shapes
Streams and References
Explorations and Patterns
Case Study: Implementing the Minority Report Mall Scene
Part One: Creating the Solution Design for the Scenario
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Conclusion
Appendix A: Scripts and Samples used in the Case Study
Appendix B: Creating a Message
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Table of Contents
Understanding Shapes, Streams, References and Explorations
Streams and References
Explorations and Patterns
dy: Implementing the Minority Report Mall Scene
Part One: Creating the Solution Design for the Scenario
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Appendix A: Scripts and Samples used in the Case Study
Appendix B: Creating a Message
ORACLE STREAM EXPLORER
Understanding Shapes, Streams, References and Explorations
Streams and References
Explorations and Patterns
dy: Implementing the Minority Report Mall Scene
Part One: Creating the Solution Design for the Scenario
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Appendix A: Scripts and Samples used in the Case Study
Appendix B: Creating a Message-Driven Bean that Greets
Understanding Shapes, Streams, References and Explorations
dy: Implementing the Minority Report Mall Scene
Part One: Creating the Solution Design for the Scenario
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Appendix A: Scripts and Samples used in the Case Study
Driven Bean that Greets
Understanding Shapes, Streams, References and Explorations
dy: Implementing the Minority Report Mall Scene
Part One: Creating the Solution Design for the Scenario
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Appendix A: Scripts and Samples used in the Case Study
Driven Bean that Greets
Understanding Shapes, Streams, References and Explorations
Part Two: Implementing the Artifacts in Oracle Stream Explorer
2
3
3
3
4
6
7
9
23
23
29
2
Introduction
Events are everywhere. From the moment we wake up until t
without our knowledge. According to any popular dictionary, an event is something that just happens. It
can be a thermostat being adjusted by a ho
store, or a car
analysis can be viewed as an event, and
events is important because they tell us what is going on and provide us with awareness of the current
situation around us and in the wider world.
The technique of analyzing event relationships and their consequences is called event processing.
Event proc
take action,
highway
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
unnoticed.
twenty
Most people think that many, if not all, industries already use event processing somehow, but the
reality is that few of them are
to allow the exchange of events between different systems. Bu
characteristics
place, only event delivery. Just to name a few,
of industries using event processing. But those are event
is part of their core business. Why then, i
failing to leverage it?
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of
event processing
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
processing app
capabilities. But this
to the person interested in doing the analysis.
2 | INTRODUCTION TO THE
Introduction
Events are everywhere. From the moment we wake up until t
without our knowledge. According to any popular dictionary, an event is something that just happens. It
can be a thermostat being adjusted by a ho
store, or a car
analysis can be viewed as an event, and
events is important because they tell us what is going on and provide us with awareness of the current
situation around us and in the wider world.
The technique of analyzing event relationships and their consequences is called event processing.
Event processing handles events while they are still in motion because that is the perfect moment to
take action, like paying attention to a low fuel warning light on your car to avoid running out of gas on a
highway. Once events happen, they become past tense and t
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
unnoticed. For this reason, event processing has never been as important as it is now, especially for
twenty-first-century enterprises.
Most people think that many, if not all, industries already use event processing somehow, but the
reality is that few of them are
to allow the exchange of events between different systems. Bu
characteristics
place, only event delivery. Just to name a few,
of industries using event processing. But those are event
is part of their core business. Why then, i
failing to leverage it?
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of
event processing
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
processing app
capabilities. But this
to the person interested in doing the analysis.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Introduction
Events are everywhere. From the moment we wake up until t
without our knowledge. According to any popular dictionary, an event is something that just happens. It
can be a thermostat being adjusted by a ho
store, or a car passing through
analysis can be viewed as an event, and
events is important because they tell us what is going on and provide us with awareness of the current
situation around us and in the wider world.
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
Once events happen, they become past tense and t
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
century enterprises.
Most people think that many, if not all, industries already use event processing somehow, but the
reality is that few of them are
to allow the exchange of events between different systems. Bu
characteristics such as loose coupling and message routing; there is no actual event pr
place, only event delivery. Just to name a few,
of industries using event processing. But those are event
is part of their core business. Why then, i
failing to leverage it?
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of
event processing technologies like
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
processing applications; that include extreme low latency and high throughput, just like
capabilities. But this comes with a high price, which is exposing the technical details of the technology
to the person interested in doing the analysis.
ORACLE STREAM EXPLORER
Events are everywhere. From the moment we wake up until t
without our knowledge. According to any popular dictionary, an event is something that just happens. It
can be a thermostat being adjusted by a ho
passing through an automated toll station on a highway. Virtually everything in the final
analysis can be viewed as an event, and
events is important because they tell us what is going on and provide us with awareness of the current
situation around us and in the wider world.
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
Once events happen, they become past tense and t
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
century enterprises.
Most people think that many, if not all, industries already use event processing somehow, but the
reality is that few of them are actually using it. Most enterprises use EDA (Event
to allow the exchange of events between different systems. Bu
such as loose coupling and message routing; there is no actual event pr
place, only event delivery. Just to name a few,
of industries using event processing. But those are event
is part of their core business. Why then, i
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of
technologies like OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
s; that include extreme low latency and high throughput, just like
comes with a high price, which is exposing the technical details of the technology
to the person interested in doing the analysis.
Events are everywhere. From the moment we wake up until t
without our knowledge. According to any popular dictionary, an event is something that just happens. It
can be a thermostat being adjusted by a homeowner, a cre
an automated toll station on a highway. Virtually everything in the final
analysis can be viewed as an event, and many types of events may be related. Paying attention to
events is important because they tell us what is going on and provide us with awareness of the current
situation around us and in the wider world.
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
Once events happen, they become past tense and t
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
Most people think that many, if not all, industries already use event processing somehow, but the
actually using it. Most enterprises use EDA (Event
to allow the exchange of events between different systems. Bu
such as loose coupling and message routing; there is no actual event pr
place, only event delivery. Just to name a few, Automat
of industries using event processing. But those are event
is part of their core business. Why then, if event processing is
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of
OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
s; that include extreme low latency and high throughput, just like
comes with a high price, which is exposing the technical details of the technology
to the person interested in doing the analysis.
Events are everywhere. From the moment we wake up until the next day, millions of events
without our knowledge. According to any popular dictionary, an event is something that just happens. It
meowner, a credit card being processed at the
an automated toll station on a highway. Virtually everything in the final
types of events may be related. Paying attention to
events is important because they tell us what is going on and provide us with awareness of the current
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
Once events happen, they become past tense and therefore become nothing more than a
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
Most people think that many, if not all, industries already use event processing somehow, but the
actually using it. Most enterprises use EDA (Event
to allow the exchange of events between different systems. But despite the usage of inherent
such as loose coupling and message routing; there is no actual event pr
utomated Trading and
of industries using event processing. But those are event-driven industries by nature; event processing
f event processing is that important, are other industries still
The main reason is likely the level of abstraction of current technologies. When compared to business
intelligence, event processing technologies offers a low level of abstraction. Even the most powerful
OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
s; that include extreme low latency and high throughput, just like
comes with a high price, which is exposing the technical details of the technology
he next day, millions of events
without our knowledge. According to any popular dictionary, an event is something that just happens. It
dit card being processed at the
an automated toll station on a highway. Virtually everything in the final
types of events may be related. Paying attention to
events is important because they tell us what is going on and provide us with awareness of the current
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
herefore become nothing more than a
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
Most people think that many, if not all, industries already use event processing somehow, but the
actually using it. Most enterprises use EDA (Event-
t despite the usage of inherent
such as loose coupling and message routing; there is no actual event pr
rading and Online Gaming
driven industries by nature; event processing
important, are other industries still
The main reason is likely the level of abstraction of current technologies. When compared to business
abstraction. Even the most powerful
OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
s; that include extreme low latency and high throughput, just like
comes with a high price, which is exposing the technical details of the technology
he next day, millions of events happen
without our knowledge. According to any popular dictionary, an event is something that just happens. It
dit card being processed at the grocery
an automated toll station on a highway. Virtually everything in the final
types of events may be related. Paying attention to
events is important because they tell us what is going on and provide us with awareness of the current
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
herefore become nothing more than a
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
Most people think that many, if not all, industries already use event processing somehow, but the
-Driven Architecture)
t despite the usage of inherent
such as loose coupling and message routing; there is no actual event processing in
aming are examples
driven industries by nature; event processing
important, are other industries still
The main reason is likely the level of abstraction of current technologies. When compared to business
abstraction. Even the most powerful
OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
s; that include extreme low latency and high throughput, just like fault-tolerant
comes with a high price, which is exposing the technical details of the technology
happen
without our knowledge. According to any popular dictionary, an event is something that just happens. It
grocery
an automated toll station on a highway. Virtually everything in the final
types of events may be related. Paying attention to
events is important because they tell us what is going on and provide us with awareness of the current
The technique of analyzing event relationships and their consequences is called event processing.
essing handles events while they are still in motion because that is the perfect moment to
like paying attention to a low fuel warning light on your car to avoid running out of gas on a
herefore become nothing more than a
record of a fact. Analyzing historical facts still has its value, but depending of the events that occurred
and the situation those events represented, it may signify a missed opportunity or a threat that went
For this reason, event processing has never been as important as it is now, especially for
Most people think that many, if not all, industries already use event processing somehow, but the
Driven Architecture)
t despite the usage of inherent EDA
ocessing in
examples
driven industries by nature; event processing
important, are other industries still
The main reason is likely the level of abstraction of current technologies. When compared to business
abstraction. Even the most powerful
OEP (Oracle Event Processing) are focused on the developer
community. In this context, OEP offers a complete set of tools to create, test, debug and deploy event
tolerant
comes with a high price, which is exposing the technical details of the technology
3
There is a clear nee
need to understand all the technical details that can confuse business users. This was the reason
Oracle launched
that allows the creation of event processing applications through an intuitive
Oracle Stream Explorer is a web
SaaS in the Oracle Cloud, ca
analyzing streams of events in real
The goal of this paper is to provide the basic information necessary to start building applications using
Oracle Stream Explorer. It will provid
step
Understanding Shapes, Streams, References and Explorations
This section will be focused in providing a
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing int
Shapes
According to many p
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
wedges, circles or
even when they are observed individually, geons also mean something because in the human brain, every geon is
matched against a structural representation of these o
shapes.
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applic
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
about a data set. It has a name, a list of attributes a
associated with a
each source
never pers
finishes all the data is discarded.
Streams and References
Events in general can be classified according to their temporal state. There is previously processed data that
represent facts from
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
the events happening right now, there is not eno
3 | INTRODUCTION TO THE
There is a clear nee
need to understand all the technical details that can confuse business users. This was the reason
Oracle launched
that allows the creation of event processing applications through an intuitive
Oracle Stream Explorer is a web
SaaS in the Oracle Cloud, ca
analyzing streams of events in real
The goal of this paper is to provide the basic information necessary to start building applications using
Oracle Stream Explorer. It will provid
step-by-step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
This section will be focused in providing a
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing int
Shapes
According to many p
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
wedges, circles or
even when they are observed individually, geons also mean something because in the human brain, every geon is
matched against a structural representation of these o
shapes.
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applic
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
about a data set. It has a name, a list of attributes a
associated with a
each source type is implemented through an
never persisted. During the analysis of
finishes all the data is discarded.
Streams and References
Events in general can be classified according to their temporal state. There is previously processed data that
represent facts from
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
the events happening right now, there is not eno
INTRODUCTION TO THE ORACLE STREAM EXPLOR
There is a clear need in the
need to understand all the technical details that can confuse business users. This was the reason
Oracle launched the new Oracle Stream Explorer
that allows the creation of event processing applications through an intuitive
Oracle Stream Explorer is a web
SaaS in the Oracle Cloud, ca
analyzing streams of events in real
The goal of this paper is to provide the basic information necessary to start building applications using
Oracle Stream Explorer. It will provid
step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
This section will be focused in providing a
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing int
According to many psychologists
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
wedges, circles or rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
matched against a structural representation of these o
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applic
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
about a data set. It has a name, a list of attributes a
associated with a source type. A source type
type is implemented through an
isted. During the analysis of
finishes all the data is discarded.
Streams and References
Events in general can be classified according to their temporal state. There is previously processed data that
represent facts from the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
the events happening right now, there is not eno
ORACLE STREAM EXPLORER
d in the industry for a product
need to understand all the technical details that can confuse business users. This was the reason
Oracle Stream Explorer
that allows the creation of event processing applications through an intuitive
Oracle Stream Explorer is a web-enabled application available to be used both on
SaaS in the Oracle Cloud, capable of providing a zero
analyzing streams of events in real-time.
The goal of this paper is to provide the basic information necessary to start building applications using
Oracle Stream Explorer. It will provide a solid introduction to the main product features and will show
step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
This section will be focused in providing a solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing int
sychologists, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
matched against a structural representation of these o
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applic
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
about a data set. It has a name, a list of attributes a
A source type determines
type is implemented through an OEP adapter. Another important aspect about shapes is that they are
isted. During the analysis of stream of
finishes all the data is discarded.
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
the events happening right now, there is not eno
for a product that provide
need to understand all the technical details that can confuse business users. This was the reason
Oracle Stream Explorer product
that allows the creation of event processing applications through an intuitive
enabled application available to be used both on
pable of providing a zero
The goal of this paper is to provide the basic information necessary to start building applications using
e a solid introduction to the main product features and will show
step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing int
, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
matched against a structural representation of these objects. In the real world, these so called geons are known as
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applic
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
about a data set. It has a name, a list of attributes and their respective data types,
determines from where the data
OEP adapter. Another important aspect about shapes is that they are
stream of events, data is brought into memory but as soon as the analysis
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
the events happening right now, there is not enough data to understand the entire context. As mentioned before, the
provides event processing power but without the
need to understand all the technical details that can confuse business users. This was the reason
product; to provide
that allows the creation of event processing applications through an intuitive
enabled application available to be used both on
pable of providing a zero-coding environment for people in
The goal of this paper is to provide the basic information necessary to start building applications using
e a solid introduction to the main product features and will show
step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
Stream Explorer spend some time reading this section before venturing into the product.
, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
bjects. In the real world, these so called geons are known as
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
it has a structural representation. Building event processing applications is no different. Every piece of data being
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
nd their respective data types,
where the data
OEP adapter. Another important aspect about shapes is that they are
events, data is brought into memory but as soon as the analysis
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
ugh data to understand the entire context. As mentioned before, the
event processing power but without the
need to understand all the technical details that can confuse business users. This was the reason
to provide business users with a platform
that allows the creation of event processing applications through an intuitive and simple user interface.
enabled application available to be used both on
coding environment for people in
The goal of this paper is to provide the basic information necessary to start building applications using
e a solid introduction to the main product features and will show
step how to develop a sample application based on an interesting case study.
Understanding Shapes, Streams, References and Explorations
solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
o the product.
, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
bjects. In the real world, these so called geons are known as
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
ations is no different. Every piece of data being
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
data set. Without creating a shape, there is no way in Oracle Stream Explorer to process any data.
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
nd their respective data types, and it is
where the data will come from, and behind the scenes
OEP adapter. Another important aspect about shapes is that they are
events, data is brought into memory but as soon as the analysis
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
ugh data to understand the entire context. As mentioned before, the
event processing power but without the
need to understand all the technical details that can confuse business users. This was the reason
ss users with a platform
and simple user interface.
enabled application available to be used both on-premise and via
coding environment for people interested in
The goal of this paper is to provide the basic information necessary to start building applications using
e a solid introduction to the main product features and will show
step how to develop a sample application based on an interesting case study.
solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
bjects. In the real world, these so called geons are known as
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
ations is no different. Every piece of data being
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
ny data.
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
it is commonly used when
from, and behind the scenes
OEP adapter. Another important aspect about shapes is that they are
events, data is brought into memory but as soon as the analysis
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
ugh data to understand the entire context. As mentioned before, the
event processing power but without the
need to understand all the technical details that can confuse business users. This was the reason
ss users with a platform
and simple user interface.
premise and via
terested in
The goal of this paper is to provide the basic information necessary to start building applications using
e a solid introduction to the main product features and will show
solid introduction to the most important artifacts created during the
development of event processing applications. It is highly recommended that first time users working with Oracle
, when human beings start learning something new, they mentally break down
objects into simple geometric forms called geons. Geons are simple 2D or 3D forms such as cylinders, bricks,
rectangles. When observed together, all geons can describe a higher geon such as a car, but
even when they are observed individually, geons also mean something because in the human brain, every geon is
bjects. In the real world, these so called geons are known as
Shapes are the fundamental building blocks in which all the objects are classified. Every object is a shape therefore
ations is no different. Every piece of data being
analyzed or helping in the analysis is a shape. You need to create a shape to provide a working representation of a
From the technical point of view, a shape in Oracle Stream Explorer is a type of artifact that provides metadata
used when
from, and behind the scenes
OEP adapter. Another important aspect about shapes is that they are
events, data is brought into memory but as soon as the analysis
Events in general can be classified according to their temporal state. There is previously processed data that
the past and there are the ongoing events that represent what is happening right now. It is
undeniable that event processing is all about discovering what is happening right now but, in most cases, with only
ugh data to understand the entire context. As mentioned before, the
4
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current str
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
may be delaye
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
are the ongoing events and they r
previously processed data, therefore they are facts from the past.
In Oracle Stream Explorer, any
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any
add more context to
are used to reference already processed data. From the temporal state perspective,
happening right now and the
Explorations and Patterns
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
artifact called an exp
Every exploration needs to be associated w
one stream must exist in the list of sources, and it must be the first
other streams
between
prior to the exploration.
The output r
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
between
4 | INTRODUCTION TO THE
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current str
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
may be delayed. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
are the ongoing events and they r
previously processed data, therefore they are facts from the past.
In Oracle Stream Explorer, any
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any
add more context to
are used to reference already processed data. From the temporal state perspective,
happening right now and the
Explorations and Patterns
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
artifact called an exp
Every exploration needs to be associated w
one stream must exist in the list of sources, and it must be the first
other streams or references, and there is no limit to the number of
between sources
prior to the exploration.
The output result of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
between sources
INTRODUCTION TO THE ORACLE STREAM EXPLOR
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current str
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
are the ongoing events and they r
previously processed data, therefore they are facts from the past.
In Oracle Stream Explorer, any
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any
add more context to a stream it is c
are used to reference already processed data. From the temporal state perspective,
happening right now and the references represents
Explorations and Patterns
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
artifact called an exploration.
Every exploration needs to be associated w
one stream must exist in the list of sources, and it must be the first
or references, and there is no limit to the number of
sources and explorations, all
prior to the exploration.
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
sources and exploration
ORACLE STREAM EXPLORER
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current str
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
are the ongoing events and they represent what is happening right now, and the details about each ship represents
previously processed data, therefore they are facts from the past.
In Oracle Stream Explorer, any unbounded sequence
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any
a stream it is called reference. References are
are used to reference already processed data. From the temporal state perspective,
references represents
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
Every exploration needs to be associated with a list of sources, which
one stream must exist in the list of sources, and it must be the first
or references, and there is no limit to the number of
and explorations, all sources
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
and explorations.
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current str
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
epresent what is happening right now, and the details about each ship represents
previously processed data, therefore they are facts from the past.
unbounded sequence of data used to represent an ongoing event is called strea
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any
alled reference. References are
are used to reference already processed data. From the temporal state perspective,
references represents facts from the past.
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
ith a list of sources, which
one stream must exist in the list of sources, and it must be the first
or references, and there is no limit to the number of
sources intended to be used in explorations need to be created in advance,
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
past events being used to provide more context to the current stream of events.
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
epresent what is happening right now, and the details about each ship represents
previously processed data, therefore they are facts from the past.
used to represent an ongoing event is called strea
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
analysis requires some of this contextual or historical data to be useful. Any static data
alled reference. References are static and the
are used to reference already processed data. From the temporal state perspective,
facts from the past.
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
ith a list of sources, which can be
one stream must exist in the list of sources, and it must be the first source
or references, and there is no limit to the number of sources used in the list.
intended to be used in explorations need to be created in advance,
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
eam of events.
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
epresent what is happening right now, and the details about each ship represents
used to represent an ongoing event is called strea
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
static data that represents data used to
static and the name came from its usage: they
are used to reference already processed data. From the temporal state perspective, the streams represent
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, pe
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
can be a stream or a reference.
source on the list. The other
used in the list.
intended to be used in explorations need to be created in advance,
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
epresent what is happening right now, and the details about each ship represents
used to represent an ongoing event is called strea
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
that represents data used to
name came from its usage: they
streams represent
During the development of event processing applications, there are a set of common activities that must be
performed to achieve some desired business goal, such as creating junctions, temporal constraints, performing
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
a reference. But, at
on the list. The other sources
Due to the relationship
intended to be used in explorations need to be created in advance,
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
analysis of historical facts still has its value, and in the world of event processing, this value is based on the idea of
For instance, imagine that many ships carrying commercial products continuously send events about their locations
and from those events you need to detect and alert when a particular ship is delayed which means product delivery
d. Assuming that the event holds at least the identifier of the ship, you would still need an external
data source to find out where the ship is coming from, registry, etc. In this context, the events coming from the ships
epresent what is happening right now, and the details about each ship represents
used to represent an ongoing event is called stream. A
stream is continuous; it never stops and provides the raw material for the event processing analysis. But most
that represents data used to
name came from its usage: they
what is
During the development of event processing applications, there are a set of common activities that must be
rforming
aggregation and filtering and customizing the output result. In most cases, these activities must be repeatedly
performed until the desired business goal is reached. In Oracle Stream Explorer, this is accomplished through an
But, at least
sources can be
Due to the relationship
intended to be used in explorations need to be created in advance,
esult of an exploration is invariably a new stream. For this reason, once the exploration is published, it
can be used in the list of sources of other explorations, just like any other stream. Figure 1 shows the relationship
5
Figure 1
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection o
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
problem instead of on the implementation details.
The usage of patterns is fairly simple: After choosing
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
export the exploration will be normally determin
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and f
As previously me
nothing more than
of an EPN. Because
5 | INTRODUCTION TO THE
Figure 1. The relationship between
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection o
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
problem instead of on the implementation details.
The usage of patterns is fairly simple: After choosing
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
» Top N: Used to obtain the first "N" events from the event stream.
» Bottom N
» Up Trend
» Down Trend
» Fluctuation
specific time window.
» Eliminate Duplicates
» Detect Duplicates
» Detect Missing Event
» W: Detects when a given field value rises and falls in "W
» Inverse W
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
export the exploration will be normally determin
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and f
As previously me
nothing more than
of an EPN. Because
INTRODUCTION TO THE ORACLE STREAM EXPLOR
. The relationship between
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection o
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
problem instead of on the implementation details.
The usage of patterns is fairly simple: After choosing
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
Used to obtain the first "N" events from the event stream.
Bottom N: Used to obtain the last "N" events from the event stream.
Up Trend: Detects when a numeric field shows a specified trend change higher in value.
Down Trend: Detects when a numeric field s
Fluctuation: Detects when a given field value changes in a specific upward or downward fashion within a
specific time window.
Eliminate Duplicates: Eliminates duplicate events in the event stream.
Duplicates: Detects when a given data field has duplicate values within a specified period of time.
Detect Missing Event: Detects when an expected event does not occur within a specific time window.
: Detects when a given field value rises and falls in "W
Inverse W: Use this pattern to detect inverse W.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
export the exploration will be normally determin
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and f
As previously mentioned, Oracle Stream Explorer has
nothing more than EPNs (Event Processing Networks)
of an EPN. Because explorations are EPNs, they can be exported
ORACLE STREAM EXPLORER
sources and explorations.
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection o
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
problem instead of on the implementation details.
The usage of patterns is fairly simple: After choosing
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
Used to obtain the first "N" events from the event stream.
: Used to obtain the last "N" events from the event stream.
: Detects when a numeric field shows a specified trend change higher in value.
: Detects when a numeric field s
: Detects when a given field value changes in a specific upward or downward fashion within a
: Eliminates duplicate events in the event stream.
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
: Detects when a given field value rises and falls in "W
: Use this pattern to detect inverse W.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
export the exploration will be normally determin
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and f
ntioned, Oracle Stream Explorer has
PNs (Event Processing Networks)
orations are EPNs, they can be exported
and explorations.
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection o
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
problem instead of on the implementation details.
The usage of patterns is fairly simple: After choosing which pattern needs to be implemented, specific data or key
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
Used to obtain the first "N" events from the event stream.
: Used to obtain the last "N" events from the event stream.
: Detects when a numeric field shows a specified trend change higher in value.
: Detects when a numeric field shows a specified trend change lower in value.
: Detects when a given field value changes in a specific upward or downward fashion within a
: Eliminates duplicate events in the event stream.
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
: Detects when a given field value rises and falls in "W
: Use this pattern to detect inverse W.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
export the exploration will be normally determined when the exploration is meant to become an OEP application;
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and f
ntioned, Oracle Stream Explorer has OEP as its foundation,
PNs (Event Processing Networks) deployed as
orations are EPNs, they can be exported
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
logic. For this reason, Oracle Stream Explorer makes available a collection of pre
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
which pattern needs to be implemented, specific data or key
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
following patterns are currently available in Oracle Stream Explorer:
Used to obtain the first "N" events from the event stream.
: Used to obtain the last "N" events from the event stream.
: Detects when a numeric field shows a specified trend change higher in value.
hows a specified trend change lower in value.
: Detects when a given field value changes in a specific upward or downward fashion within a
: Eliminates duplicate events in the event stream.
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
: Detects when a given field value rises and falls in "W" fashion over a specified time window.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
ed when the exploration is meant to become an OEP application;
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
technical details such as integration, security, performance, scalability and fault
as its foundation,
deployed as OEP applications. Figure 2 show
orations are EPNs, they can be exported from Oracle Stream Explorer
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
f pre-built explorations called patterns.
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
which pattern needs to be implemented, specific data or key
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
: Detects when a numeric field shows a specified trend change higher in value.
hows a specified trend change lower in value.
: Detects when a given field value changes in a specific upward or downward fashion within a
: Eliminates duplicate events in the event stream.
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
" fashion over a specified time window.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
ed when the exploration is meant to become an OEP application;
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
ault-tolerance.
as its foundation, and the generated explorations are
OEP applications. Figure 2 show
Oracle Stream Explorer
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
built explorations called patterns.
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
which pattern needs to be implemented, specific data or key
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
: Detects when a numeric field shows a specified trend change higher in value.
hows a specified trend change lower in value.
: Detects when a given field value changes in a specific upward or downward fashion within a
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
" fashion over a specified time window.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
ed when the exploration is meant to become an OEP application;
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
erated explorations are
OEP applications. Figure 2 shows an example
Oracle Stream Explorer and imported again
Working with explorations can be very time consuming, especially if the scenario being explored requires complex
built explorations called patterns.
Patterns are generic solutions to recurring common problems, and they help to keep the focus on the business
which pattern needs to be implemented, specific data or key
fields will be required, and after providing all of them the entire exploration will be automatically generated. The
: Detects when a given field value changes in a specific upward or downward fashion within a
: Detects when a given data field has duplicate values within a specified period of time.
: Detects when an expected event does not occur within a specific time window.
Once an exploration is built it can be used as a source for another exploration or it can be exported. The decision to
ed when the exploration is meant to become an OEP application;
originally created in Oracle Stream Explorer to reach the business goal but to be later enhanced in OEP to handle
erated explorations are
s an example
and imported again
6
into a development environment
JAR file is generated with
CQL (Continuous Query Language) statements.
Figure 2
Case Study: Implementing the Minority Report Mall Scene
In July of 2002, 2
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
The film revolves around the story of a spe
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
faces a challenge since he is n
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
of its own eyes and have then surgically exchanged wit
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
saying "Hello Mr. Yakamoto, welcome back to the GAP".
What this scene taught us is that the technology of sensors combined with event processing can pr
individualized experience for
customer's
retinas and send the results to a capable event processing technology for processing. The event processing
c
after recogniz
This scenario would not be so challenging if it was not for the fact that
custom message while the person is near the store. If the current business intelligence techniques were used in this
scenario, the events
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
6 | INTRODUCTION TO THE
into a development environment
JAR file is generated with
CQL (Continuous Query Language) statements.
Figure 2. Example of an EPN generated from the implementation of an exploration.
Case Study: Implementing the Minority Report Mall Scene
In July of 2002, 2
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
The film revolves around the story of a spe
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
faces a challenge since he is n
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
of its own eyes and have then surgically exchanged wit
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
saying "Hello Mr. Yakamoto, welcome back to the GAP".
What this scene taught us is that the technology of sensors combined with event processing can pr
individualized experience for
customer's personal preferences
retinas and send the results to a capable event processing technology for processing. The event processing
comes into play when it has to,
after recognizing
This scenario would not be so challenging if it was not for the fact that
custom message while the person is near the store. If the current business intelligence techniques were used in this
scenario, the events
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
INTRODUCTION TO THE ORACLE STREAM EXPLOR
into a development environment
JAR file is generated with common EPN artifacts such as e
CQL (Continuous Query Language) statements.
Example of an EPN generated from the implementation of an exploration.
Case Study: Implementing the Minority Report Mall Scene
In July of 2002, 20th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
The film revolves around the story of a spe
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
faces a challenge since he is n
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
of its own eyes and have then surgically exchanged wit
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
saying "Hello Mr. Yakamoto, welcome back to the GAP".
What this scene taught us is that the technology of sensors combined with event processing can pr
individualized experience for customers,
personal preferences
retinas and send the results to a capable event processing technology for processing. The event processing
omes into play when it has to,
ing who the person is, greeting them with a custom message.
This scenario would not be so challenging if it was not for the fact that
custom message while the person is near the store. If the current business intelligence techniques were used in this
scenario, the events from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
ORACLE STREAM EXPLORER
into a development environment such as Oracle Fusion Middleware JDeveloper. When the exploration is ex
common EPN artifacts such as e
CQL (Continuous Query Language) statements.
Example of an EPN generated from the implementation of an exploration.
Case Study: Implementing the Minority Report Mall Scene
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
The film revolves around the story of a special police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
faces a challenge since he is now the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
of its own eyes and have then surgically exchanged wit
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
saying "Hello Mr. Yakamoto, welcome back to the GAP".
What this scene taught us is that the technology of sensors combined with event processing can pr
customers, reacting appropriately when ne
personal preferences. The sensor in this
retinas and send the results to a capable event processing technology for processing. The event processing
omes into play when it has to, from a stream of multiple
who the person is, greeting them with a custom message.
This scenario would not be so challenging if it was not for the fact that
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
Oracle Fusion Middleware JDeveloper. When the exploration is ex
common EPN artifacts such as event types, adapters, caches, channels, processors and
CQL (Continuous Query Language) statements.
Example of an EPN generated from the implementation of an exploration.
Case Study: Implementing the Minority Report Mall Scene
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
cial police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
ow the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
of its own eyes and have then surgically exchanged with a pair of eyes that belonged to a deceased person known
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
saying "Hello Mr. Yakamoto, welcome back to the GAP".
What this scene taught us is that the technology of sensors combined with event processing can pr
reacting appropriately when ne
sensor in this context are the eye scanners that continuously read people's
retinas and send the results to a capable event processing technology for processing. The event processing
m a stream of multiple events;
who the person is, greeting them with a custom message.
This scenario would not be so challenging if it was not for the fact that
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
Oracle Fusion Middleware JDeveloper. When the exploration is ex
vent types, adapters, caches, channels, processors and
Example of an EPN generated from the implementation of an exploration.
Case Study: Implementing the Minority Report Mall Scene
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
cial police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
ow the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
h a pair of eyes that belonged to a deceased person known
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised
What this scene taught us is that the technology of sensors combined with event processing can pr
reacting appropriately when needed and
are the eye scanners that continuously read people's
retinas and send the results to a capable event processing technology for processing. The event processing
events; detect which one of them is near the store and,
who the person is, greeting them with a custom message.
This scenario would not be so challenging if it was not for the fact that the event processing system
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
Oracle Fusion Middleware JDeveloper. When the exploration is ex
vent types, adapters, caches, channels, processors and
Case Study: Implementing the Minority Report Mall Scene
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
cial police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
murders before they happen, deploying the police force in preemptive fashion to arrest the future murderers. In the
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesti
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
ow the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
h a pair of eyes that belonged to a deceased person known
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
situation when, looking for some new clothes, he enters in a GAP store and it is surprised with a greeting message
What this scene taught us is that the technology of sensors combined with event processing can pr
eded and with information about the
are the eye scanners that continuously read people's
retinas and send the results to a capable event processing technology for processing. The event processing
detect which one of them is near the store and,
the event processing system
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
Oracle Fusion Middleware JDeveloper. When the exploration is exported, a
vent types, adapters, caches, channels, processors and
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
cial police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
arrest the future murderers. In the
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
John Anderton; a police captain responsible for the PreCrime program. Besides presenting a very interesting story,
the film also has lots of scenes in which high technology is used, revealing how the near future could be.
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
ow the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
h a pair of eyes that belonged to a deceased person known
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
with a greeting message
What this scene taught us is that the technology of sensors combined with event processing can provide
with information about the
are the eye scanners that continuously read people's
retinas and send the results to a capable event processing technology for processing. The event processing
detect which one of them is near the store and,
the event processing system must trigger the
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
ported, a
vent types, adapters, caches, channels, processors and
0th Century Fox's studios released Minority Report. Directed by Steven Spielberg and starring Tom
Cruise as the main character, the film rapidly became one of the most successful science fiction movies ever made.
cial police program called PreCrime. Using three mutated humans stored
in a special chamber where they experience precognitive visions, the PreCrime program is designed to stop
arrest the future murderers. In the
film, the PreCrime program is established in Washington D.C in the year of 2050, and Tom Cruise plays the role of
ng story,
In one of those scenes, John Anderton needs to get in into the chamber where the mutated humans are, but he
ow the main suspect of a future crime. In the year of 2050, nearly all public places
have eye scanners that can identify every citizen. So in order to not be apprehended he has no choice but to get rid
h a pair of eyes that belonged to a deceased person known
as Yakamoto. After the surgery and already in possession of Mr. Yakamoto eyes, he finds himself into an odd
with a greeting message
ovide an
with information about the
are the eye scanners that continuously read people's
retinas and send the results to a capable event processing technology for processing. The event processing part
detect which one of them is near the store and,
must trigger the
custom message while the person is near the store. If the current business intelligence techniques were used in this
from the scanned people would be stored in a staging database where a scheduled job would
load the last "N" scans that were near the store, possibly an hour, day or even a week later. To make sense, the
7
custom message must be delivered while the person is
a few seconds.
This is when the technology of event processing comes into play making it feasible to actually build a system that
could make this type of science fiction scene possible. Even
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
near the stor
order to make the implementation feasible, the following assumptions will be assumed:
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
Part One: Creating the Solution Design for the Scenario
When designing a solution that is going to be implemented in Oracle St
artifacts that wi
all data that is going to flow through the event processing application, the first step of the
of a conceptual model, separating the shapes in
are going to be needed. Figure 3 shows the conceptual model of this case study.
Figure 3
In order to classify each shape according to its
associated with
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from the Oracle Stream Explorer perspective, they
invariably a new stream.
instead of
7 | INTRODUCTION TO THE
custom message must be delivered while the person is
a few seconds.
This is when the technology of event processing comes into play making it feasible to actually build a system that
could make this type of science fiction scene possible. Even
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
near the store to the moment in which he or she
order to make the implementation feasible, the following assumptions will be assumed:
» Each eye scan event holds the person retina and his or her location.
» All customer information is available through a database table.
» The custom greetings are avai
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
Part One: Creating the Solution Design for the Scenario
When designing a solution that is going to be implemented in Oracle St
artifacts that will be required
all data that is going to flow through the event processing application, the first step of the
of a conceptual model, separating the shapes in
are going to be needed. Figure 3 shows the conceptual model of this case study.
Figure 3. Shapes in the
In order to classify each shape according to its
associated with ongoing events are stereotyped as streams and shapes
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from the Oracle Stream Explorer perspective, they
invariably a new stream.
instead of a stream, because in the final analysis every exploration is a stream.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
custom message must be delivered while the person is
This is when the technology of event processing comes into play making it feasible to actually build a system that
could make this type of science fiction scene possible. Even
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
e to the moment in which he or she
order to make the implementation feasible, the following assumptions will be assumed:
Each eye scan event holds the person retina and his or her location.
All customer information is available through a database table.
The custom greetings are avai
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
Part One: Creating the Solution Design for the Scenario
When designing a solution that is going to be implemented in Oracle St
ll be required, starting with the necessary
all data that is going to flow through the event processing application, the first step of the
of a conceptual model, separating the shapes in
are going to be needed. Figure 3 shows the conceptual model of this case study.
Shapes in the conceptual model of the case study.
In order to classify each shape according to its
ongoing events are stereotyped as streams and shapes
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from the Oracle Stream Explorer perspective, they
invariably a new stream. For this reason, it would not be considered a mistake using the stereotype exploration
stream, because in the final analysis every exploration is a stream.
ORACLE STREAM EXPLORER
custom message must be delivered while the person is
This is when the technology of event processing comes into play making it feasible to actually build a system that
could make this type of science fiction scene possible. Even
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
e to the moment in which he or she
order to make the implementation feasible, the following assumptions will be assumed:
Each eye scan event holds the person retina and his or her location.
All customer information is available through a database table.
The custom greetings are available through a database table.
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
Part One: Creating the Solution Design for the Scenario
When designing a solution that is going to be implemented in Oracle St
, starting with the necessary
all data that is going to flow through the event processing application, the first step of the
of a conceptual model, separating the shapes in
are going to be needed. Figure 3 shows the conceptual model of this case study.
conceptual model of the case study.
In order to classify each shape according to its
ongoing events are stereotyped as streams and shapes
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from the Oracle Stream Explorer perspective, they
r this reason, it would not be considered a mistake using the stereotype exploration
stream, because in the final analysis every exploration is a stream.
custom message must be delivered while the person is in close proximity the store, a situation that may only last for
This is when the technology of event processing comes into play making it feasible to actually build a system that
could make this type of science fiction scene possible. Event processing enables the analysis of what is happening
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
e to the moment in which he or she is recognized to finally, receive a custom greeting mess
order to make the implementation feasible, the following assumptions will be assumed:
Each eye scan event holds the person retina and his or her location.
All customer information is available through a database table.
lable through a database table.
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
Part One: Creating the Solution Design for the Scenario
When designing a solution that is going to be implemented in Oracle St
, starting with the necessary shapes. Considering that
all data that is going to flow through the event processing application, the first step of the
of a conceptual model, separating the shapes into two categories: the shapes already available and the shapes that
are going to be needed. Figure 3 shows the conceptual model of this case study.
conceptual model of the case study.
In order to classify each shape according to its usage, the conceptual model
ongoing events are stereotyped as streams and shapes
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from the Oracle Stream Explorer perspective, they derive from explorations and the output result of an explorat
r this reason, it would not be considered a mistake using the stereotype exploration
stream, because in the final analysis every exploration is a stream.
in close proximity the store, a situation that may only last for
This is when the technology of event processing comes into play making it feasible to actually build a system that
t processing enables the analysis of what is happening
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be develop
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
is recognized to finally, receive a custom greeting mess
order to make the implementation feasible, the following assumptions will be assumed:
Each eye scan event holds the person retina and his or her location.
All customer information is available through a database table.
lable through a database table.
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
When designing a solution that is going to be implemented in Oracle Stream Explorer, you need to detail all the
shapes. Considering that
all data that is going to flow through the event processing application, the first step of the
two categories: the shapes already available and the shapes that
are going to be needed. Figure 3 shows the conceptual model of this case study.
the conceptual model
ongoing events are stereotyped as streams and shapes associated with
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from explorations and the output result of an explorat
r this reason, it would not be considered a mistake using the stereotype exploration
stream, because in the final analysis every exploration is a stream.
in close proximity the store, a situation that may only last for
This is when the technology of event processing comes into play making it feasible to actually build a system that
t processing enables the analysis of what is happening
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
scene from the Minority Report film will be the case study that is going to be developed in this section. Using Oracle
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
is recognized to finally, receive a custom greeting mess
order to make the implementation feasible, the following assumptions will be assumed:
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
ream Explorer, you need to detail all the
shapes. Considering that shapes are the representation of
all data that is going to flow through the event processing application, the first step of the solution design is t
two categories: the shapes already available and the shapes that
are going to be needed. Figure 3 shows the conceptual model of this case study.
the conceptual model may include stereotypes. Shapes
associated with already processed data are
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from explorations and the output result of an explorat
r this reason, it would not be considered a mistake using the stereotype exploration
stream, because in the final analysis every exploration is a stream.
in close proximity the store, a situation that may only last for
This is when the technology of event processing comes into play making it feasible to actually build a system that
t processing enables the analysis of what is happening
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
ed in this section. Using Oracle
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
is recognized to finally, receive a custom greeting mess
The development of this case study will be divided in two parts. The first part will cover all the details related to the
solution design, where all the necessary artifacts will be identified. The second part will cover the development of
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
ream Explorer, you need to detail all the
shapes are the representation of
solution design is t
two categories: the shapes already available and the shapes that
include stereotypes. Shapes
already processed data are
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from explorations and the output result of an explorat
r this reason, it would not be considered a mistake using the stereotype exploration
in close proximity the store, a situation that may only last for
This is when the technology of event processing comes into play making it feasible to actually build a system that
t processing enables the analysis of what is happening
right now, handling events while they are still ongoing and detecting complex relationships between them. The mall
ed in this section. Using Oracle
Stream Explorer, this case study is going to be built through explorations, covering the detection of when a person is
is recognized to finally, receive a custom greeting message. In
The development of this case study will be divided in two parts. The first part will cover all the details related to the
development of
those artifacts in Oracle Stream Explorer, showing how to create them and how to handle the configuration details.
ream Explorer, you need to detail all the
shapes are the representation of
solution design is to build
two categories: the shapes already available and the shapes that
include stereotypes. Shapes
already processed data are
stereotyped as references. All the shapes found in the needed category may be stereotyped as streams because,
from explorations and the output result of an exploration is
r this reason, it would not be considered a mistake using the stereotype exploration
8
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
the causality rel
Figure 4
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
in advance how many explorations will need to be
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
a reasonable level.
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ex
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eye scan stream ne
find out who the person near the store is.
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
Explorer provides a high l
any,
8 | INTRODUCTION TO THE
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
the causality relationships.
Figure 4. Conceptual model refined to include the causality relationships.
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
in advance how many explorations will need to be
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
a reasonable level.
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
exploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eye scan stream ne
find out who the person near the store is.
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
Explorer provides a high l
any, software development
INTRODUCTION TO THE ORACLE STREAM EXPLOR
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
ationships.
Conceptual model refined to include the causality relationships.
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
in advance how many explorations will need to be
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
a reasonable level.
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eye scan stream needs to be correlated with the scanEntry attribute present in the customer reference, in order to
find out who the person near the store is.
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
Explorer provides a high level of abstraction in regards to implementation details. For this reason, people with few
software development skills can use the product directly, being responsible for the end
ORACLE STREAM EXPLORER
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Conceptual model refined to include the causality relationships.
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
in advance how many explorations will need to be
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
find out who the person near the store is.
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few
skills can use the product directly, being responsible for the end
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Conceptual model refined to include the causality relationships.
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
in advance how many explorations will need to be built is important because it can reveal how much development
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few
skills can use the product directly, being responsible for the end
The next step is the inclusion of the causality relationships between all the shapes, emp
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Conceptual model refined to include the causality relationships.
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
built is important because it can reveal how much development
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a be
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
Each exploration must have a meaningful name. As a rule of thumb, consider that each expl
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few
skills can use the product directly, being responsible for the end
The next step is the inclusion of the causality relationships between all the shapes, emphasizing how the processing
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
built is important because it can reveal how much development
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
causalities need to be implemented in different explorations, it is considered a best practice to break down the
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
Each exploration must have a meaningful name. As a rule of thumb, consider that each exploration must be named
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few
skills can use the product directly, being responsible for the end-to
hasizing how the processing
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the executio
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
built is important because it can reveal how much development
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
st practice to break down the
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
oration must be named
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
The target audience of this description will be the person responsible for the implementation of the artifacts in Or
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few
to-end implementation.
hasizing how the processing
of one or more shapes will result in the creation of a new shape. In this context, the term processing refers to the
usage of any activity of junction, temporal constraint, aggregation or filtering, performed during the execution of an
exploration. It is also important that during this step the attributes of each shape be included, just like the mapping
between the attributes of the shapes available and the shapes needed. Figure 4 shows the conceptual model with
Each lane shown on figure 4 represents an exploration that needs to be built in Oracle Stream Explorer. Finding out
built is important because it can reveal how much development
effort the scenario will demand from the implementation perspective. While there is no rule that dictates that
st practice to break down the
business problem into smaller explorations, promoting better reuse of the artifacts and moving the complexity toward
oration must be named
with the shape name that it intends to create. If business rules need to be applied during the execution of the
exploration, it is considered a best practice to provide a description of these business rules. For instance, the first
ploration has as business rule the utilization of the latitude and longitude attributes present in the eye scan stream
to detect if the customer is near the store or not. Also considered as a business rule, the eyeScan attribute from the
eds to be correlated with the scanEntry attribute present in the customer reference, in order to
The target audience of this description will be the person responsible for the implementation of the artifacts in Oracle
Stream Explorer, which can be the same person that builds the conceptual model or it can be someone different. In
most cases, the same person that builds the conceptual model also implements the artifacts since Oracle Stream
evel of abstraction in regards to implementation details. For this reason, people with few, if
end implementation.
9
If different people need to cooperate to impl
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or,
Figure 5 shows the final version of the conceptual model.
Figure 5
The conceptual model detailed in figure 5 provides a solid foundat
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
commonly found in this type of solution.
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Before start
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
of
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
be found in the Appendix A.
Considering that two
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
9 | INTRODUCTION TO THE
f different people need to cooperate to impl
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or,
Figure 5 shows the final version of the conceptual model.
Figure 5. Final version of the conceptual model with the description of the business rules.
The conceptual model detailed in figure 5 provides a solid foundat
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
commonly found in this type of solution.
Part Two: Implementing the Artifacts in Oracle Stream Explorer
Before starting the imp
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
of-the-box supported databases are Oracle, SQL
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
be found in the Appendix A.
Considering that two
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
INTRODUCTION TO THE ORACLE STREAM EXPLOR
f different people need to cooperate to impl
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or,
Figure 5 shows the final version of the conceptual model.
Final version of the conceptual model with the description of the business rules.
The conceptual model detailed in figure 5 provides a solid foundat
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
commonly found in this type of solution.
Part Two: Implementing the Artifacts in Oracle Stream Explorer
the implementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
box supported databases are Oracle, SQL
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
be found in the Appendix A.
Considering that two sources that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ORACLE STREAM EXPLORER
f different people need to cooperate to implement the scenario, providing a description of the business rules can
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or,
Figure 5 shows the final version of the conceptual model.
Final version of the conceptual model with the description of the business rules.
The conceptual model detailed in figure 5 provides a solid foundat
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
commonly found in this type of solution.
Part Two: Implementing the Artifacts in Oracle Stream Explorer
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
box supported databases are Oracle, SQL
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ement the scenario, providing a description of the business rules can
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or,
Figure 5 shows the final version of the conceptual model.
Final version of the conceptual model with the description of the business rules.
The conceptual model detailed in figure 5 provides a solid foundat
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
Part Two: Implementing the Artifacts in Oracle Stream Explorer
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
box supported databases are Oracle, SQL Server 2005 and Derby. Other JDBC compliant databases can be
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ement the scenario, providing a description of the business rules can
help in the communication throughout the entire project implementation. This is particularly important
solution designer and the implementer belong to different companies or, when
Final version of the conceptual model with the description of the business rules.
The conceptual model detailed in figure 5 provides a solid foundation for the implementation of the case study. In
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
Part Two: Implementing the Artifacts in Oracle Stream Explorer
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
Server 2005 and Derby. Other JDBC compliant databases can be
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementati
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ement the scenario, providing a description of the business rules can
help in the communication throughout the entire project implementation. This is particularly important
en they are geographically separated.
Final version of the conceptual model with the description of the business rules.
ion for the implementation of the case study. In
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution des
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
Server 2005 and Derby. Other JDBC compliant databases can be
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
create and populate the two database tables that will be used during the implementation. This SQL script can be
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ement the scenario, providing a description of the business rules can
help in the communication throughout the entire project implementation. This is particularly important when
they are geographically separated.
ion for the implementation of the case study. In
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
business scenarios. The steps shown in this section are not intended to be used as a solution design methodology.
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out
Server 2005 and Derby. Other JDBC compliant databases can be
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
on. This SQL script can be
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection p
ement the scenario, providing a description of the business rules can
when the
they are geographically separated.
ion for the implementation of the case study. In
most cases, the creation of the conceptual model will be elaborated mentally during the implementation of the
ign methodology.
Instead, they show how a typical event processing application should be designed and what kinds of concerns are
lementation of the artifacts in Oracle Stream Explorer, it is necessary to make sure that some
prerequisites are satisfied. First, a running database with DDL and DML permissions needs to be available. The out-
Server 2005 and Derby. Other JDBC compliant databases can be
used, but deployment of their JDBC drivers may be necessary. Secondly, an SQL script needs to be executed to
on. This SQL script can be
found in the Appendix A of this paper. Finally, a CSV file will be used during the implementation of the eye scan
stream artifact. Sample data from the content of this file, just like the URL to download its complete version can also
that act as references to a database table will need to be created in Oracle Stream
Explorer, a database connection pool needs to be set up. In order to be able to set up a database connection pool,
10
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL:
address or hostname where the server is running and port is the HTTP
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
Figure 6
Appropriate
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
where Oracle St
tabs. To start the creation of the databas
10 | INTRODUCTION TO THE
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL:
address or hostname where the server is running and port is the HTTP
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
Figure 6. Oracle Event Processing Visualizer lo
Appropriate credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
where Oracle St
tabs. To start the creation of the databas
INTRODUCTION TO THE ORACLE STREAM EXPLOR
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL:
address or hostname where the server is running and port is the HTTP
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
. Oracle Event Processing Visualizer lo
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
where Oracle Stream Explorer is running. This will open the window where all the server details will be available via
tabs. To start the creation of the databas
ORACLE STREAM EXPLORER
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL:
address or hostname where the server is running and port is the HTTP
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
. Oracle Event Processing Visualizer login screen.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
tabs. To start the creation of the database connection pool, click in the
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL:
address or hostname where the server is running and port is the HTTP
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
gin screen.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
e connection pool, click in the
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
up and running, open a web browser and access the following URL: http://host:port/wlevs
address or hostname where the server is running and port is the HTTP-enabled port configured for external access.
Figure 6 shows the login screen of the Oracle Event Processing Visualizer.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
e connection pool, click in the DataSource
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
http://host:port/wlevs. In this UR
enabled port configured for external access.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
DataSource tab as shown in figure 7.
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
. In this URL, host is the IP
enabled port configured for external access.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
as shown in figure 7.
administrative access to the Oracle Event Processing Visualizer will be necessary. With the Oracle Stream Explorer
L, host is the IP
enabled port configured for external access.
credentials will be needed to access this console. If you do not have these credentials, ask your Oracle
Stream Explorer administrator to create them for you. Once logged in to the console, access the server instance
ream Explorer is running. This will open the window where all the server details will be available via
11
Figure 7
The
button to create a new database connection pool. The new data source creation wizard will then start. Ente
value "jdbc/minorityReport" for both the
Figure 8
Set the
here is only reading the data available in the database table, and setting this field to none decreases the overhead
incurred in the database. Next, click on the
the database.
11 | INTRODUCTION TO THE
Figure 7. Starting the database
The DataSource
button to create a new database connection pool. The new data source creation wizard will then start. Ente
value "jdbc/minorityReport" for both the
Figure 8. Starting the creation of the database connection pool.
Set the Global Transaction Protocol
here is only reading the data available in the database table, and setting this field to none decreases the overhead
incurred in the database. Next, click on the
the database.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Starting the database connection pool creation in the console.
DataSource tab shows all the da
button to create a new database connection pool. The new data source creation wizard will then start. Ente
value "jdbc/minorityReport" for both the
Starting the creation of the database connection pool.
Global Transaction Protocol
here is only reading the data available in the database table, and setting this field to none decreases the overhead
incurred in the database. Next, click on the
ORACLE STREAM EXPLORER
connection pool creation in the console.
tab shows all the database connection pools already
button to create a new database connection pool. The new data source creation wizard will then start. Ente
value "jdbc/minorityReport" for both the Name and the
Starting the creation of the database connection pool.
Global Transaction Protocol field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
incurred in the database. Next, click on the Global Tra
connection pool creation in the console.
tabase connection pools already
button to create a new database connection pool. The new data source creation wizard will then start. Ente
and the JNDI Name
Starting the creation of the database connection pool.
field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
Global Transaction Protocol
connection pool creation in the console.
tabase connection pools already created for that server. Click on the
button to create a new database connection pool. The new data source creation wizard will then start. Ente
JNDI Name fields, as shown in figure 8.
field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
nsaction Protocol tab to enter the JDBC connection details of
created for that server. Click on the
button to create a new database connection pool. The new data source creation wizard will then start. Ente
fields, as shown in figure 8.
field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
tab to enter the JDBC connection details of
created for that server. Click on the
button to create a new database connection pool. The new data source creation wizard will then start. Enter with the
fields, as shown in figure 8.
field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
tab to enter the JDBC connection details of
created for that server. Click on the Add
r with the
field to "None" as shown in figure 8. This is very important because the intention
here is only reading the data available in the database table, and setting this field to none decreases the overhead
tab to enter the JDBC connection details of
12
Figure 9
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
the configuration of the JDBC con
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
Stream Explorer application using the following URL:
access the Oracle Event Processing Visualizer, and the same credentials t
Figure 10 shows the login screen of the Oracle Stream Explorer.
Figure 10
12 | INTRODUCTION TO THE
Figure 9. Entering the JDBC connection details of the database.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
the configuration of the JDBC con
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
Stream Explorer application using the following URL:
access the Oracle Event Processing Visualizer, and the same credentials t
Figure 10 shows the login screen of the Oracle Stream Explorer.
Figure 10. Login screen of the Oracle Stream Explorer.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Entering the JDBC connection details of the database.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
the configuration of the JDBC con
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
Stream Explorer application using the following URL:
access the Oracle Event Processing Visualizer, and the same credentials t
Figure 10 shows the login screen of the Oracle Stream Explorer.
Login screen of the Oracle Stream Explorer.
ORACLE STREAM EXPLORER
Entering the JDBC connection details of the database.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
the configuration of the JDBC connection details of your database is finished, click in the
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
Stream Explorer application using the following URL:
access the Oracle Event Processing Visualizer, and the same credentials t
Figure 10 shows the login screen of the Oracle Stream Explorer.
Login screen of the Oracle Stream Explorer.
Entering the JDBC connection details of the database.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
nection details of your database is finished, click in the
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
Stream Explorer application using the following URL: http://host:port/sx
access the Oracle Event Processing Visualizer, and the same credentials t
Figure 10 shows the login screen of the Oracle Stream Explorer.
Login screen of the Oracle Stream Explorer.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the
nection details of your database is finished, click in the
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artif
http://host:port/sx. Those are the same host and port used to
access the Oracle Event Processing Visualizer, and the same credentials that were used before should work here.
Figure 10 shows the login screen of the Oracle Stream Explorer.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
implementation. Regardless of which type of database is being used, set the value of the Use XA
nection details of your database is finished, click in the Save
creation of the database connection pool. Figure 9 shows an example of this configuration.
With the database connection pool properly created, the development of the artifacts can start. Access the Oracle
. Those are the same host and port used to
hat were used before should work here.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
Use XA field to false. Once
Save button to finish the
acts can start. Access the Oracle
. Those are the same host and port used to
hat were used before should work here.
Enter the requested JDBC connections details of the database that contains the two tables needed for this
field to false. Once
button to finish the
acts can start. Access the Oracle
. Those are the same host and port used to
hat were used before should work here.
13
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
of the business do
without setting a business domain tag,
The catalog screen is where artifacts in Oracl
previously created in the
explorations, streams and references. To create artifacts, you mu
New Item
Figure 11
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
field enter the v
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
13 | INTRODUCTION TO THE
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
of the business do
without setting a business domain tag,
The catalog screen is where artifacts in Oracl
previously created in the
explorations, streams and references. To create artifacts, you mu
New Item combo box, just as shown in figure 11.
Figure 11. Requesting the creation of new artifacts in Oracle Stream Explorer.
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
field enter the value "CustomerData". Provide the tag "customer" in the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
of the business domain and also drives you directly to the catalog screen. The catalog screen can also be accessed
without setting a business domain tag,
The catalog screen is where artifacts in Oracl
previously created in the main
explorations, streams and references. To create artifacts, you mu
combo box, just as shown in figure 11.
Requesting the creation of new artifacts in Oracle Stream Explorer.
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
alue "CustomerData". Provide the tag "customer" in the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
ORACLE STREAM EXPLORER
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
without setting a business domain tag, via the Catalog
The catalog screen is where artifacts in Oracle Stream Explorer are created and maintained. It displays all artifacts
main table on the center of the screen,
explorations, streams and references. To create artifacts, you mu
combo box, just as shown in figure 11.
Requesting the creation of new artifacts in Oracle Stream Explorer.
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
alue "CustomerData". Provide the tag "customer" in the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
Catalog button, found on the top right corner of the home screen.
e Stream Explorer are created and maintained. It displays all artifacts
le on the center of the screen,
explorations, streams and references. To create artifacts, you mu
combo box, just as shown in figure 11.
Requesting the creation of new artifacts in Oracle Stream Explorer.
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
alue "CustomerData". Provide the tag "customer" in the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
button, found on the top right corner of the home screen.
e Stream Explorer are created and maintained. It displays all artifacts
le on the center of the screen, allowing the filtering of specific artifacts such as
explorations, streams and references. To create artifacts, you must request the creation of a new one in the
Requesting the creation of new artifacts in Oracle Stream Explorer.
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
alue "CustomerData". Provide the tag "customer" in the Tags field and from the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The ho
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
button, found on the top right corner of the home screen.
e Stream Explorer are created and maintained. It displays all artifacts
allowing the filtering of specific artifacts such as
st request the creation of a new one in the
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
field and from the
box, choose the option "Database Table". Figure 12 shows the first step of this wizard.
After getting authenticated to Oracle Stream Explorer, the home screen will be displayed. The home screen provides
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
button, found on the top right corner of the home screen.
e Stream Explorer are created and maintained. It displays all artifacts
allowing the filtering of specific artifacts such as
st request the creation of a new one in the
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the
field and from the Source Type
me screen provides
a welcome message and displays a set of fancy images, each one inside a square that represents the common
business domains where event processing applications can be used. Clicking into one of these images sets the tag
main and also drives you directly to the catalog screen. The catalog screen can also be accessed
button, found on the top right corner of the home screen.
e Stream Explorer are created and maintained. It displays all artifacts
allowing the filtering of specific artifacts such as
st request the creation of a new one in the Create
The first artifact that will be created is the customer reference. Request the creation of a new reference as shown in
figure 11, selecting the option "Reference" in the list. The new reference creation wizard will then start. In the Name
Source Type combo
14
Figure 12
Click in
database connection pool that will be used to connect to the database
choose the option "jdbc/minorityReport" as
Figure 13
In the third and last step,
the
button to finish the new reference creation wizard.
14 | INTRODUCTION TO THE
Figure 12. First step of the new reference creation wizard.
Click in the Next
database connection pool that will be used to connect to the database
choose the option "jdbc/minorityReport" as
Figure 13. Second step of the new reference creation wizard.
In the third and last step,
the Select Shape
button to finish the new reference creation wizard.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
First step of the new reference creation wizard.
Next button to proceed to the second step of the wizard.
database connection pool that will be used to connect to the database
choose the option "jdbc/minorityReport" as
Second step of the new reference creation wizard.
In the third and last step, the wizard will ask
Select Shape combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
button to finish the new reference creation wizard.
ORACLE STREAM EXPLORER
First step of the new reference creation wizard.
button to proceed to the second step of the wizard.
database connection pool that will be used to connect to the database
choose the option "jdbc/minorityReport" as shown in figure 13. Click in the
Second step of the new reference creation wizard.
the wizard will ask for the name of the database table that the reference should point to
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
button to finish the new reference creation wizard.
First step of the new reference creation wizard.
button to proceed to the second step of the wizard.
database connection pool that will be used to connect to the database
shown in figure 13. Click in the
Second step of the new reference creation wizard.
for the name of the database table that the reference should point to
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
button to finish the new reference creation wizard.
button to proceed to the second step of the wizard. The wizard will
database connection pool that will be used to connect to the database server.
shown in figure 13. Click in the Next
for the name of the database table that the reference should point to
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
The wizard will ask for the name of the
server. In the Data Source Name
Next button to proceed to the third step.
for the name of the database table that the reference should point to
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
ask for the name of the
Data Source Name combo box,
button to proceed to the third step.
for the name of the database table that the reference should point to
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the
ask for the name of the
combo box,
button to proceed to the third step.
for the name of the database table that the reference should point to. In
combo box, choose the option "CUSTOMER_DATA" as shown in figure 14. Click in the Create
15
Figure 14
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
reference, set the
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
references.
Figure 15
Now it
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
creati
wizard will then start.
15 | INTRODUCTION TO THE
Figure 14. Third step of the new reference creation wizard.
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
reference, set the
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
references.
Figure 15. Oracle Stream Explorer showing the two cre
Now it is necessary the creation of a
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
creation of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
wizard will then start.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Third step of the new reference creation wizard.
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
reference, set the Name field to "CustomGreeting" and
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
Oracle Stream Explorer showing the two cre
is necessary the creation of a
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
wizard will then start.
ORACLE STREAM EXPLORER
Third step of the new reference creation wizard.
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
ield to "CustomGreeting" and
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
Oracle Stream Explorer showing the two cre
is necessary the creation of a third source,
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
Third step of the new reference creation wizard.
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
ield to "CustomGreeting" and the shape selected must be the
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
Oracle Stream Explorer showing the two created references.
source, but this time it will be a stream instead of a reference. This stream
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
the shape selected must be the
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
ated references.
but this time it will be a stream instead of a reference. This stream
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
the shape selected must be the "CUSTOM_GREETING"
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
but this time it will be a stream instead of a reference. This stream
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
"CUSTOM_GREETING"
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
but this time it will be a stream instead of a reference. This stream
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
After the completion of the wizard, a new shape named "CustomerData" should be listed in the table in the center of
the screen. Now, the same steps must be followed to create the second reference. During the creation of the second
"CUSTOM_GREETING"
database table. Figure 15 shows what should be exhibited in the catalog screen after the creation of these two
but this time it will be a stream instead of a reference. This stream
represents the incoming flow of eye scans containing the person's retina reading and its location. Request the
on of a new stream as shown in figure 11, selecting the option "Stream" in the list. The new stream creation
16
In the
Source Type
unchecked, since all the explorations are going to be created in the sequence. Click in the
the second step of the wizard. Since the sel
source path of the CSV file. Click in the
Explorer.
Click the
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
finish the new stream creation wizard.
Figure 16
All
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
that detects whe
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
figure 11, selecting the option "Exploration
In the
from the
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
exploration as shown in figure 18.
16 | INTRODUCTION TO THE
In the Name field enter the value "EyeScanStream". Provide the tag "customer" in the
Source Type co
unchecked, since all the explorations are going to be created in the sequence. Click in the
the second step of the wizard. Since the sel
source path of the CSV file. Click in the
Explorer.
Click the Next button to proceed to the third step. In the
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
finish the new stream creation wizard.
Figure 16. All the three steps from the new stream creation wizard.
All sources needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
that detects whe
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
figure 11, selecting the option "Exploration
In the Name field enter the value "All Customers Near the Store". Provide the tag "customer" in the
from the Source
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
exploration as shown in figure 18.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
field enter the value "EyeScanStream". Provide the tag "customer" in the
combo box, select the option "CSV File". Leave the
unchecked, since all the explorations are going to be created in the sequence. Click in the
the second step of the wizard. Since the sel
source path of the CSV file. Click in the
button to proceed to the third step. In the
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
finish the new stream creation wizard.
All the three steps from the new stream creation wizard.
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
that detects when some person is
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
figure 11, selecting the option "Exploration
field enter the value "All Customers Near the Store". Provide the tag "customer" in the
Source combo box, choose the option "EyeScanStream", just as sho
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
exploration as shown in figure 18.
ORACLE STREAM EXPLORER
field enter the value "EyeScanStream". Provide the tag "customer" in the
mbo box, select the option "CSV File". Leave the
unchecked, since all the explorations are going to be created in the sequence. Click in the
the second step of the wizard. Since the selected source type was set to CSV, the wizard will then ask for the
source path of the CSV file. Click in the Upload File
button to proceed to the third step. In the
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
finish the new stream creation wizard.
All the three steps from the new stream creation wizard.
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
n some person is walking near the store, also identifying
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
figure 11, selecting the option "Exploration" in the list. The new exploration creation wizard will then start.
field enter the value "All Customers Near the Store". Provide the tag "customer" in the
combo box, choose the option "EyeScanStream", just as sho
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
exploration as shown in figure 18.
field enter the value "EyeScanStream". Provide the tag "customer" in the
mbo box, select the option "CSV File". Leave the
unchecked, since all the explorations are going to be created in the sequence. Click in the
ected source type was set to CSV, the wizard will then ask for the
Upload File button to select and upload the CSV file to Oracle Stream
button to proceed to the third step. In the Name
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
All the three steps from the new stream creation wizard.
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
near the store, also identifying
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
" in the list. The new exploration creation wizard will then start.
field enter the value "All Customers Near the Store". Provide the tag "customer" in the
combo box, choose the option "EyeScanStream", just as sho
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
field enter the value "EyeScanStream". Provide the tag "customer" in the
mbo box, select the option "CSV File". Leave the Create Exploration with this Source
unchecked, since all the explorations are going to be created in the sequence. Click in the
ected source type was set to CSV, the wizard will then ask for the
button to select and upload the CSV file to Oracle Stream
Name field enter the value "EyeScanStream" again. During
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make s
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
near the store, also identifying them
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
" in the list. The new exploration creation wizard will then start.
field enter the value "All Customers Near the Store". Provide the tag "customer" in the
combo box, choose the option "EyeScanStream", just as sho
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
field enter the value "EyeScanStream". Provide the tag "customer" in the Tags
Create Exploration with this Source
unchecked, since all the explorations are going to be created in the sequence. Click in the Next
ected source type was set to CSV, the wizard will then ask for the
button to select and upload the CSV file to Oracle Stream
ld enter the value "EyeScanStream" again. During
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
order to be able to present them in the third step of the wizard. For this reason, make sure if the
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
them against the customer reference. This
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
" in the list. The new exploration creation wizard will then start.
field enter the value "All Customers Near the Store". Provide the tag "customer" in the
combo box, choose the option "EyeScanStream", just as shown in figure 17. Click in the
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
Tags field and from the
Create Exploration with this Source checkbox
Next button to proceed to
ected source type was set to CSV, the wizard will then ask for the
button to select and upload the CSV file to Oracle Stream
ld enter the value "EyeScanStream" again. During
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
ure if the Manual Mapping
field is selected, and if the attributes shown matches with what is shown in figure 16. Click in the Create button to
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
against the customer reference. This
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
" in the list. The new exploration creation wizard will then start.
field enter the value "All Customers Near the Store". Provide the tag "customer" in the Tags field and
wn in figure 17. Click in the
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
field and from the
checkbox
button to proceed to
ected source type was set to CSV, the wizard will then ask for the
button to select and upload the CSV file to Oracle Stream
ld enter the value "EyeScanStream" again. During
the file upload process, the wizard parses the CSV file trying to detect the attribute names and their data types, in
Manual Mapping
button to
needed to start the implementation of the case study are now created. Now it is time to start the creation
of the explorations, the artifacts that will provide the expected behavior. The first exploration to be built is the one
against the customer reference. This
exploration will be named "All Customers Near the Store". Request the creation of a new exploration as shown in
field and
wn in figure 17. Click in the Create
button to finish the new exploration creation wizard. The exploration editor will then open with the newly created
17
Figure 17
The first thing noticed in the exploration editor is that the
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
the exploration is
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
coming from the sources.
Figure 18
17 | INTRODUCTION TO THE
Figure 17. Setting the source in the new exploration creation wizar
The first thing noticed in the exploration editor is that the
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
the exploration is
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
coming from the sources.
Figure 18. Newly created exploration presenting in real
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Setting the source in the new exploration creation wizar
The first thing noticed in the exploration editor is that the
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
the exploration is changed, these sections are also updated, in real
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
coming from the sources.
created exploration presenting in real
ORACLE STREAM EXPLORER
Setting the source in the new exploration creation wizar
The first thing noticed in the exploration editor is that the
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
changed, these sections are also updated, in real
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
created exploration presenting in real
Setting the source in the new exploration creation wizard.
The first thing noticed in the exploration editor is that the Live Output Stream
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
changed, these sections are also updated, in real
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
created exploration presenting in real-time the data coming from the sources.
Live Output Stream and the
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
changed, these sections are also updated, in real-time, helping the user to have a better view of
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
time the data coming from the sources.
and the Charts
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
time, helping the user to have a better view of
what the output result looks like. Figure 18 shows the newly created exploration presenting in real
time the data coming from the sources.
sections shows in real
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
time, helping the user to have a better view of
what the output result looks like. Figure 18 shows the newly created exploration presenting in real-time the data
sections shows in real-
time the results coming from the sources, which in this case is the eye scan stream. This is interesting because as
time, helping the user to have a better view of
time the data
18
Since the exploration need
need to be created. In the
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
are looking for. For this exploration, the filters w
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
Figure 19
After setting these filters, the
the criteria, and the
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every eve
the customer reference that holds the details of the customer. The first step is
exploration to include another shape. Click on the
the
Correlations
Click in the
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
Live Output Stream
18 | INTRODUCTION TO THE
Since the exploration need
need to be created. In the
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
are looking for. For this exploration, the filters w
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
Figure 19. Filtering the person's location in Oracle Stream Explorer.
After setting these filters, the
the criteria, and the
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every eve
the customer reference that holds the details of the customer. The first step is
exploration to include another shape. Click on the
the drop down list. When two or more shapes are used as sources in an exploration, a new section called
Correlations start
Click in the Add a Correlation
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
Live Output Stream
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Since the exploration needs to detect only persons that are near the store, filters to pick up only the ones o
need to be created. In the Filters
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
are looking for. For this exploration, the filters w
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
Filtering the person's location in Oracle Stream Explorer.
After setting these filters, the Live
the criteria, and the Charts section will then start to show a more flat graph, since the live output stream now
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every eve
the customer reference that holds the details of the customer. The first step is
exploration to include another shape. Click on the
drop down list. When two or more shapes are used as sources in an exploration, a new section called
starts to appear in the exploration editor.
Add a Correlation link to create a new correlation between the sources. From the left
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
Live Output Stream section will show attributes from both sources, as shown in figure 20.
ORACLE STREAM EXPLORER
to detect only persons that are near the store, filters to pick up only the ones o
ilters section, click in the
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
are looking for. For this exploration, the filters w
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
Filtering the person's location in Oracle Stream Explorer.
ive Output Stream
section will then start to show a more flat graph, since the live output stream now
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every eve
the customer reference that holds the details of the customer. The first step is
exploration to include another shape. Click on the
drop down list. When two or more shapes are used as sources in an exploration, a new section called
pear in the exploration editor.
link to create a new correlation between the sources. From the left
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
section will show attributes from both sources, as shown in figure 20.
to detect only persons that are near the store, filters to pick up only the ones o
section, click in the Add a Filter
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
are looking for. For this exploration, the filters will guarantee that only persons in which its location has the latitude
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
Filtering the person's location in Oracle Stream Explorer.
tream section will automatically discard any event that does not satisfies
section will then start to show a more flat graph, since the live output stream now
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every eve
the customer reference that holds the details of the customer. The first step is
exploration to include another shape. Click on the Sources section and select the "CustomerData" reference from
drop down list. When two or more shapes are used as sources in an exploration, a new section called
pear in the exploration editor.
link to create a new correlation between the sources. From the left
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
section will show attributes from both sources, as shown in figure 20.
to detect only persons that are near the store, filters to pick up only the ones o
Add a Filter link. A new filter entry will be created where you
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
ill guarantee that only persons in which its location has the latitude
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
shows these two filters being created in Oracle Stream Explorer.
will automatically discard any event that does not satisfies
section will then start to show a more flat graph, since the live output stream now
presents only approximate values for the latitude and longitude attributes.
In order to identify who each person is, it is necessary that every event from the eye scan str
the customer reference that holds the details of the customer. The first step is
section and select the "CustomerData" reference from
drop down list. When two or more shapes are used as sources in an exploration, a new section called
link to create a new correlation between the sources. From the left
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
section will show attributes from both sources, as shown in figure 20.
to detect only persons that are near the store, filters to pick up only the ones o
link. A new filter entry will be created where you
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
ill guarantee that only persons in which its location has the latitude
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "
will automatically discard any event that does not satisfies
section will then start to show a more flat graph, since the live output stream now
from the eye scan str
the customer reference that holds the details of the customer. The first step is to change the
section and select the "CustomerData" reference from
drop down list. When two or more shapes are used as sources in an exploration, a new section called
link to create a new correlation between the sources. From the left
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
section will show attributes from both sources, as shown in figure 20.
to detect only persons that are near the store, filters to pick up only the ones of
link. A new filter entry will be created where you
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
ill guarantee that only persons in which its location has the latitude
attribute greater than or equals to "20.00" and the longitude attribute is less than or equals to "-60.00". Figure 19
will automatically discard any event that does not satisfies
section will then start to show a more flat graph, since the live output stream now
from the eye scan stream be correlated with
change the Sources section of the
section and select the "CustomerData" reference from
drop down list. When two or more shapes are used as sources in an exploration, a new section called
link to create a new correlation between the sources. From the left side choose the
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
interest
link. A new filter entry will be created where you
can choose which attribute to handle, which operator to use and set which value restricts the output result that you
ill guarantee that only persons in which its location has the latitude
60.00". Figure 19
will automatically discard any event that does not satisfies
section will then start to show a more flat graph, since the live output stream now
eam be correlated with
section of the
section and select the "CustomerData" reference from
drop down list. When two or more shapes are used as sources in an exploration, a new section called
side choose the
"eyeScan" attribute and from the right side choose the "scan_entry" attribute. Once the correlation is configured, the
19
Figure 20
In order to be considered complete, explorations need to provide the expected output result since once published,
they
output result via
and "gender", also re
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
shown in figure 21.
Figure 21
19 | INTRODUCTION TO THE
Figure 20. Result of the
In order to be considered complete, explorations need to provide the expected output result since once published,
they may be used in the list of sources of other explorations. Oracl
output result via
and "gender", also re
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
shown in figure 21.
Figure 21. Customizing the output result of an exploration before publishing it.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Result of the event correlation between the eye scan stream and the customer reference.
In order to be considered complete, explorations need to provide the expected output result since once published,
be used in the list of sources of other explorations. Oracl
output result via the Properties
and "gender", also re-ordering them to be the first and second attributes of the propertie
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
shown in figure 21.
Customizing the output result of an exploration before publishing it.
ORACLE STREAM EXPLORER
event correlation between the eye scan stream and the customer reference.
In order to be considered complete, explorations need to provide the expected output result since once published,
be used in the list of sources of other explorations. Oracl
link. Access the
ordering them to be the first and second attributes of the propertie
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
Customizing the output result of an exploration before publishing it.
event correlation between the eye scan stream and the customer reference.
In order to be considered complete, explorations need to provide the expected output result since once published,
be used in the list of sources of other explorations. Oracl
Access the Properties link
ordering them to be the first and second attributes of the propertie
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
Customizing the output result of an exploration before publishing it.
event correlation between the eye scan stream and the customer reference.
In order to be considered complete, explorations need to provide the expected output result since once published,
be used in the list of sources of other explorations. Oracle Stream Explorer allows the mo
link and leave selected only the attributes "last_name"
ordering them to be the first and second attributes of the propertie
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
Customizing the output result of an exploration before publishing it.
event correlation between the eye scan stream and the customer reference.
In order to be considered complete, explorations need to provide the expected output result since once published,
e Stream Explorer allows the mo
selected only the attributes "last_name"
ordering them to be the first and second attributes of the properties section. After this, double
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
In order to be considered complete, explorations need to provide the expected output result since once published,
e Stream Explorer allows the modification of the
selected only the attributes "last_name"
s section. After this, double
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
In order to be considered complete, explorations need to provide the expected output result since once published,
dification of the
selected only the attributes "last_name"
s section. After this, double-
click the "last_name" attribute to provide a custom display name for it, setting its value to "customerName", as
20
The explor
labeled
possible action that can be performed in the
Figure 22
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
as shown in figure 11, s
start.
In the
and from the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
showing in the
and select the "CustomGreeting" reference from the drop down list. Once the
the exploration editor, click on the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
is configured, the
20 | INTRODUCTION TO THE
The exploration is now ready to be published. In the top right corner of the exploration editor there is a button
labeled Actions
possible action that can be performed in the
Figure 22. Publishing the exploration to be available as a stream.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
as shown in figure 11, s
start.
In the Name field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
and from the Source
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
showing in the Live Output Stream
and select the "CustomGreeting" reference from the drop down list. Once the
the exploration editor, click on the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
is configured, the
INTRODUCTION TO THE ORACLE STREAM EXPLOR
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
possible action that can be performed in the
Publishing the exploration to be available as a stream.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
as shown in figure 11, selecting the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
Source combo box, choose the option "All Customers Near the Store". Click the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
Live Output Stream
and select the "CustomGreeting" reference from the drop down list. Once the
the exploration editor, click on the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
is configured, the Live Output Stream
ORACLE STREAM EXPLORER
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
possible action that can be performed in the exploration. Click in the
Publishing the exploration to be available as a stream.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
combo box, choose the option "All Customers Near the Store". Click the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
Live Output Stream section the out
and select the "CustomGreeting" reference from the drop down list. Once the
the exploration editor, click on the Add a Correlation
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
Live Output Stream section will show attributes from both sources, as
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
exploration. Click in the
Publishing the exploration to be available as a stream.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
combo box, choose the option "All Customers Near the Store". Click the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
section the output result from the first exploration. Click in the
and select the "CustomGreeting" reference from the drop down list. Once the
Add a Correlation link to create a
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
section will show attributes from both sources, as
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
exploration. Click in the Publish button just as shown in figure 22.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
combo box, choose the option "All Customers Near the Store". Click the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
put result from the first exploration. Click in the
and select the "CustomGreeting" reference from the drop down list. Once the Correlations
link to create a new correlation between the sources. From the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
section will show attributes from both sources, as
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
button just as shown in figure 22.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the
combo box, choose the option "All Customers Near the Store". Click the
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
put result from the first exploration. Click in the
Correlations section starts to appear in
new correlation between the sources. From the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
section will show attributes from both sources, as shown in figure 23.
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
button just as shown in figure 22.
With the first exploration properly published, the second exploration can start to be built. The second exploration wi
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
field enter the value "Greetings for All Customers Near". Provide the tag "customer" in the Tags
combo box, choose the option "All Customers Near the Store". Click the Create button to finish
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
put result from the first exploration. Click in the Sources
section starts to appear in
new correlation between the sources. From the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
shown in figure 23.
ation is now ready to be published. In the top right corner of the exploration editor there is a button
that once clicked opens a suspended menu with other buttons, each one of them related to a
button just as shown in figure 22.
With the first exploration properly published, the second exploration can start to be built. The second exploration will
create greetings for each customer near the store, using the person gender to detect which appropriate greeting to
use. This exploration will be named "Greetings for All Customers Near". Request the creation of a new exploration
electing the option "Exploration" in the list. The new exploration creation wizard will then
Tags field
button to finish
the new exploration creation wizard. The exploration editor will then open with the newly created exploration,
section
section starts to appear in
new correlation between the sources. From the
left side choose the "gender" attribute and from the right side choose the "gender" attribute too. Once the correlation
21
Figure 23
To
back
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
Access the
"custom_message", also re
Still in the
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
just as shown in figure 24.
21 | INTRODUCTION TO THE
Figure 23. Second exploration with the correlation already in place.
To avoid unnecessary greetings for the
back to the GAP" greeting is used. In the
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
Access the Properties
"custom_message", also re
Still in the Properti
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
just as shown in figure 24.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Second exploration with the correlation already in place.
avoid unnecessary greetings for the
to the GAP" greeting is used. In the
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
roperties link and leave selected only the attributes "ice_break_message", "customerName" and the
"custom_message", also re-ordering them to be the first, second and the third attributes of the pr
roperties link, provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
just as shown in figure 24.
ORACLE STREAM EXPLORER
Second exploration with the correlation already in place.
avoid unnecessary greetings for the customers, a filter needs to be created to establish that only the "Welc
to the GAP" greeting is used. In the Filters
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
and leave selected only the attributes "ice_break_message", "customerName" and the
ordering them to be the first, second and the third attributes of the pr
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
Second exploration with the correlation already in place.
customers, a filter needs to be created to establish that only the "Welc
ilters section, click in the
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
and leave selected only the attributes "ice_break_message", "customerName" and the
ordering them to be the first, second and the third attributes of the pr
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
customers, a filter needs to be created to establish that only the "Welc
section, click in the Add a Filter
setting the value of the "greeting_code" attribute to be equals to "WBACK".
Just like it was done with the first exploration, the output result of this second exploration need
and leave selected only the attributes "ice_break_message", "customerName" and the
ordering them to be the first, second and the third attributes of the pr
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
customers, a filter needs to be created to establish that only the "Welc
Add a Filter link to create a new filter
Just like it was done with the first exploration, the output result of this second exploration need
and leave selected only the attributes "ice_break_message", "customerName" and the
ordering them to be the first, second and the third attributes of the pr
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
customers, a filter needs to be created to establish that only the "Welc
link to create a new filter
Just like it was done with the first exploration, the output result of this second exploration needs to be customized.
and leave selected only the attributes "ice_break_message", "customerName" and the
ordering them to be the first, second and the third attributes of the properties section.
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
customers, a filter needs to be created to establish that only the "Welcome
link to create a new filter entry,
tomized.
and leave selected only the attributes "ice_break_message", "customerName" and the
operties section.
provide custom display names for each one of these three attributes, setting the first
attribute to "iceBreakMessage", the second attribute to "customerName" and the third attribute to "customMessage",
22
Figure 24
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
Stream Explorer detecting in real
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
salutations
However, while it is interesting to have the ou
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows th
external systems via the
Explorer:
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
the configured target. This means that
immediately without
situation became relevant, allowing a truly event
F
MDB (Message
via the
of a target of this type, as shown in figure 25.
22 | INTRODUCTION TO THE
Figure 24. Second exploration with its output result properly customized.
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
Stream Explorer detecting in real
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
salutations such as "Mr." and "Mrs." before saying their names.
However, while it is interesting to have the ou
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows th
external systems via the
Explorer:
» CSV File
» HTTP Publisher
» Event-Driven Network
» Java Message Service
» REST Endpoints
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
the configured target. This means that
immediately without
situation became relevant, allowing a truly event
For instance, imagine that a voice
MDB (Message-
via the Java Speech API
of a target of this type, as shown in figure 25.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Second exploration with its output result properly customized.
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
Stream Explorer detecting in real
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
such as "Mr." and "Mrs." before saying their names.
However, while it is interesting to have the ou
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows th
external systems via the Configure a Target
CSV File: Write the output results to a CSV file, allowing
HTTP Publisher: Publishes the output results to an outbound HTTP channel from OEP.
Driven Network: Send the output results to SOA Suite EDN
Java Message Service: Creates a
REST Endpoints: Performs a HTTP
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
the configured target. This means that
immediately without any delays. This approach allows that actions can be taken in the exact moment in which the
situation became relevant, allowing a truly event
or instance, imagine that a voice
-Driven Bean) to consume messages from the JMS destination and execute the greetings speech
Java Speech API. The output results from the exploration could be easily sent via JMS by the configuration
of a target of this type, as shown in figure 25.
ORACLE STREAM EXPLORER
Second exploration with its output result properly customized.
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
Stream Explorer detecting in real-time when someone is near the store and
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
such as "Mr." and "Mrs." before saying their names.
However, while it is interesting to have the output result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows th
Configure a Target button. The following targets are currently supported in Oracle Stream
: Write the output results to a CSV file, allowing
: Publishes the output results to an outbound HTTP channel from OEP.
: Send the output results to SOA Suite EDN
: Creates a javax.jms.MapMessage
: Performs a HTTP POST with the output results into a REST endpoint.
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
the configured target. This means that once a new output result is available i
delays. This approach allows that actions can be taken in the exact moment in which the
situation became relevant, allowing a truly event
or instance, imagine that a voice-enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
of a target of this type, as shown in figure 25.
Second exploration with its output result properly customized.
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
n someone is near the store and
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
such as "Mr." and "Mrs." before saying their names.
tput result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows th
button. The following targets are currently supported in Oracle Stream
: Write the output results to a CSV file, allowing
: Publishes the output results to an outbound HTTP channel from OEP.
: Send the output results to SOA Suite EDN
javax.jms.MapMessage
POST with the output results into a REST endpoint.
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
once a new output result is available i
delays. This approach allows that actions can be taken in the exact moment in which the
situation became relevant, allowing a truly event-driven approach.
enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
Second exploration with its output result properly customized.
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
n someone is near the store and greeti
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
such as "Mr." and "Mrs." before saying their names.
tput result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
greeting notification. Fortunately, Oracle Stream Explorer allows that output results from explorations to be sent to
button. The following targets are currently supported in Oracle Stream
: Write the output results to a CSV file, allowing the contents to be app
: Publishes the output results to an outbound HTTP channel from OEP.
: Send the output results to SOA Suite EDN-enabled subscriber.
javax.jms.MapMessage and send it to a JMS
POST with the output results into a REST endpoint.
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
once a new output result is available i
delays. This approach allows that actions can be taken in the exact moment in which the
driven approach.
enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
greeting them with a custom message.
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
tput result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
at output results from explorations to be sent to
button. The following targets are currently supported in Oracle Stream
contents to be appended.
: Publishes the output results to an outbound HTTP channel from OEP.
enabled subscriber.
and send it to a JMS destination.
POST with the output results into a REST endpoint.
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
once a new output result is available in Oracle Stream Explorer, it is
delays. This approach allows that actions can be taken in the exact moment in which the
enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
ng them with a custom message.
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
tput result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
at output results from explorations to be sent to
button. The following targets are currently supported in Oracle Stream
ended.
: Publishes the output results to an outbound HTTP channel from OEP.
enabled subscriber.
destination.
POST with the output results into a REST endpoint.
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
n Oracle Stream Explorer, it is
delays. This approach allows that actions can be taken in the exact moment in which the
enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
Publish this second exploration as shown in figure 22. The case study is now completely implemented, with Oracle
ng them with a custom message.
By using the person's gender to differentiate if it is a man or a woman, the automated approach also uses English
tput result correctly shown in the Oracle Stream Explorer exploration
editor, this can be considered useless if this output result could not be sent to an external system that handles the
at output results from explorations to be sent to
button. The following targets are currently supported in Oracle Stream
It is important to note that regardless of the chosen target type, the generated output results are sent continuously to
n Oracle Stream Explorer, it is sent
delays. This approach allows that actions can be taken in the exact moment in which the
enabled system implemented in Java listen greeting requests via JMS, using a
Driven Bean) to consume messages from the JMS destination and execute the greetings speech
. The output results from the exploration could be easily sent via JMS by the configuration
23
Figure 25
The end
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
applications. Howeve
Speech API to synthesize the greeting using a human voice.
Conclusion
Can you imagine yourself driving your car by only looking
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
Oracle Stream Explorer is a web
to provide a tool for business users
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
to develop event processing
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and a
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
unchanged.
--------------------------------------------------------
--
--------------------------------------------------------
23 | INTRODUCTION TO THE
Figure 25. Configuration of a JMS
The end-to-end implementation of a voice
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
applications. Howeve
Speech API to synthesize the greeting using a human voice.
Conclusion
Can you imagine yourself driving your car by only looking
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
Oracle Stream Explorer is a web
to provide a tool for business users
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
to develop event processing
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and a
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
unchanged.
--------------------------------------------------------
-- DDL for the table CUSTOMER_DATA
--------------------------------------------------------
CREATE TABLE "CUSTOMER_DATA"
( "CUSTOMER_ID" INTEGER NOT NULL PRIMARY KEY,
"SCAN_ENTRY" VARCHAR2(255) NOT NULL,
"FIRST_NAME" VARCHAR2(20) NOT NULL,
"LAST_NAME" VARCHAR2(20) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL
);
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Configuration of a JMS-based target to send th
end implementation of a voice
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
applications. However, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
Speech API to synthesize the greeting using a human voice.
Can you imagine yourself driving your car by only looking
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
Oracle Stream Explorer is a web
to provide a tool for business users
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
to develop event processing-based applications through a step
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and a
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
--------------------------------------------------------
DDL for the table CUSTOMER_DATA
--------------------------------------------------------
CREATE TABLE "CUSTOMER_DATA"
"CUSTOMER_ID" INTEGER NOT NULL PRIMARY KEY,
AN_ENTRY" VARCHAR2(255) NOT NULL,
"FIRST_NAME" VARCHAR2(20) NOT NULL,
"LAST_NAME" VARCHAR2(20) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL
ORACLE STREAM EXPLORER
based target to send th
end implementation of a voice-enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
Speech API to synthesize the greeting using a human voice.
Can you imagine yourself driving your car by only looking
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
Oracle Stream Explorer is a web-based application that leverages the capabilities found in Oracle Event Processing
to provide a tool for business users to analyze streams o
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
based applications through a step
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and a
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
--------------------------------------------------------
DDL for the table CUSTOMER_DATA
--------------------------------------------------------
CREATE TABLE "CUSTOMER_DATA"
"CUSTOMER_ID" INTEGER NOT NULL PRIMARY KEY,
AN_ENTRY" VARCHAR2(255) NOT NULL,
"FIRST_NAME" VARCHAR2(20) NOT NULL,
"LAST_NAME" VARCHAR2(20) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL
based target to send the output results.
enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
Speech API to synthesize the greeting using a human voice.
Can you imagine yourself driving your car by only looking in the rear
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
based application that leverages the capabilities found in Oracle Event Processing
analyze streams of events in real
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
based applications through a step
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and a
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
--------------------------------------------------------
--------------------------------------------------------
"CUSTOMER_ID" INTEGER NOT NULL PRIMARY KEY,
AN_ENTRY" VARCHAR2(255) NOT NULL,
"FIRST_NAME" VARCHAR2(20) NOT NULL,
"LAST_NAME" VARCHAR2(20) NOT NULL,
e output results.
enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
the rear-mirror? This is how
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
based application that leverages the capabilities found in Oracle Event Processing
f events in real-time, empowering them to gain insight and
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
based applications through a step-by-step implementation of
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
The script shown in the listing 1 creates these two database tables and also populates it with sample data. The
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
--------------------------------------------------------
--------------------------------------------------------
"CUSTOMER_ID" INTEGER NOT NULL PRIMARY KEY,
enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
mirror? This is how many
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
important opportunities or letting threats take advantage of our lack of awareness.
based application that leverages the capabilities found in Oracle Event Processing
time, empowering them to gain insight and
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
step implementation of a case stud
Appendix A: Scripts and Samples used in the Case Study
In order to be able to implement the case study presented in this paper, two database tables need to be created.
lso populates it with sample data. The
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names
enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
many enterprises run their
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
based application that leverages the capabilities found in Oracle Event Processing
time, empowering them to gain insight and
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
a case study.
In order to be able to implement the case study presented in this paper, two database tables need to be created.
lso populates it with sample data. The
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
adjusts in the script can be performed, as long that the structure of the tables and the column names remains
enabled system using Java is definitely out of the scope of this paper,
which focuses on explaining the fundamentals of Oracle Stream Explorer and how to build event processing
r, the Appendix B of this paper contains an implementation of a MDB that leverages the Java
enterprises run their
business today, trying to get insight using traditional data warehouse and business intelligence technologies.
Looking at past information does not always provide the ability to react appropriately; this in turn leads to missing
based application that leverages the capabilities found in Oracle Event Processing
time, empowering them to gain insight and
take appropriate actions when needed. This paper explained the fundamentals of Oracle Stream Explorer and how
In order to be able to implement the case study presented in this paper, two database tables need to be created.
lso populates it with sample data. The
script was successfully tested against the Oracle database, but it may work well in other databases. If necessary,
remains
24
--------------------------------------------------------
--
--------------------------------------------------------
--------------------------------------------------------
--
--------------------------------------------------------
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN
(1,'CB281A82
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(2,'B7C841D6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(3,'0028CF68
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(4,'B5C9DA94
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(5,'1EB943F4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,F
(6,'B71E107F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(7,'AF25D32D
Insert
(8,'EA712817
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(9,'27AFC1DC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(10,'45D8A1B8
Insert into CUSTOMER_DATA
(11,'3FAB4B01
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(12,'F6C6D0BC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(13,'096458CC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
(14,'95005563
24 | INTRODUCTION TO THE
--------------------------------------------------------
-- DDL for the table CUSTOM_GREETING
--------------------------------------------------------
CREATE TABLE "CUSTOM_GREETING"
( "GREETING_ID" INTEGER NOT NULL PRIMARY KEY,
"GREETING_CODE" VARCHAR2(5) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL,
"ICE_BREAK_MESSAGE" VARCHAR2(50) NOT NUL
"CUSTOM_MESSAGE" VARCHAR2(100) NOT NULL
);
--------------------------------------------------------
-- Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN
(1,'CB281A82-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(2,'B7C841D6-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(3,'0028CF68-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(4,'B5C9DA94-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(5,'1EB943F4-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,F
(6,'B71E107F-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(7,'AF25D32D-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(8,'EA712817-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(9,'27AFC1DC-
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(10,'45D8A1B8
Insert into CUSTOMER_DATA
(11,'3FAB4B01
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(12,'F6C6D0BC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(13,'096458CC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
(14,'95005563
INTRODUCTION TO THE ORACLE STREAM EXPLOR
--------------------------------------------------------
DDL for the table CUSTOM_GREETING
--------------------------------------------------------
CREATE TABLE "CUSTOM_GREETING"
"GREETING_ID" INTEGER NOT NULL PRIMARY KEY,
"GREETING_CODE" VARCHAR2(5) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL,
"ICE_BREAK_MESSAGE" VARCHAR2(50) NOT NUL
"CUSTOM_MESSAGE" VARCHAR2(100) NOT NULL
--------------------------------------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN
-EBF9-247F-8D4C
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-DC7D-0C32-1402
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-2264-D591-C3F8
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-6AF7-9BFF-19F5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-C7B3-29D8-D0D3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,F
-DD89-76EC-497B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-0870-7C79-5ECE
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-9B59-042A-F952
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-74D6-D988-9E7E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(10,'45D8A1B8-72DB-A6EE-7725
Insert into CUSTOMER_DATA
(11,'3FAB4B01-AA8C-CA14-03B7
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(12,'F6C6D0BC-5772-97BA-2EB3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(13,'096458CC-3F58-0CFC-08B9
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
(14,'95005563-5DFB-2F6D-5B12
ORACLE STREAM EXPLORER
--------------------------------------------------------
DDL for the table CUSTOM_GREETING
--------------------------------------------------------
CREATE TABLE "CUSTOM_GREETING"
"GREETING_ID" INTEGER NOT NULL PRIMARY KEY,
"GREETING_CODE" VARCHAR2(5) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL,
"ICE_BREAK_MESSAGE" VARCHAR2(50) NOT NUL
"CUSTOM_MESSAGE" VARCHAR2(100) NOT NULL
--------------------------------------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN
8D4C-632F2CD4DC9E','Gregory','Berger','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1402-58E0E06F0C74','George','Griffith
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C3F8-EECF78E3635F','Damian','Key','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
19F5-9EECC4797417','Scott','Bridges','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D0D3-042507CCDD50','Colton','Kinney','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,F
497B-9F3BB81CC790','Brody','Goodman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5ECE-A8D88E63A099','Simon','Park','M');
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F952-4BD9B2621FE7','Addison','Medina','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9E7E-35522FD65CAC','Charles','Sutton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7725-4AA8F379FA2E','Vaughan','Lawson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
03B7-69F75670D0C2','Felix','Haley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2EB3-E0A6DE91D64B','M
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
08B9-08E2134AFCB7','Brent','Faulkner','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
5B12-1BB0DEE71492','Kasimir','Fitzgerald','F');
--------------------------------------------------------
DDL for the table CUSTOM_GREETING
--------------------------------------------------------
"GREETING_ID" INTEGER NOT NULL PRIMARY KEY,
"GREETING_CODE" VARCHAR2(5) NOT NULL,
"GENDER" VARCHAR2(1) NOT NULL,
"ICE_BREAK_MESSAGE" VARCHAR2(50) NOT NULL,
"CUSTOM_MESSAGE" VARCHAR2(100) NOT NULL
--------------------------------------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
632F2CD4DC9E','Gregory','Berger','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
58E0E06F0C74','George','Griffith
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EECF78E3635F','Damian','Key','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9EECC4797417','Scott','Bridges','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
042507CCDD50','Colton','Kinney','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,F
9F3BB81CC790','Brody','Goodman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A8D88E63A099','Simon','Park','M');
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4BD9B2621FE7','Addison','Medina','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35522FD65CAC','Charles','Sutton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4AA8F379FA2E','Vaughan','Lawson','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
69F75670D0C2','Felix','Haley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E0A6DE91D64B','Mason','Mclean','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
08E2134AFCB7','Brent','Faulkner','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
1BB0DEE71492','Kasimir','Fitzgerald','F');
--------------------------------------------------------
--------------------------------------------------------
"GREETING_ID" INTEGER NOT NULL PRIMARY KEY,
L,
--------------------------------------------------------
--------------------------------------------------------
_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
632F2CD4DC9E','Gregory','Berger','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
58E0E06F0C74','George','Griffith','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EECF78E3635F','Damian','Key','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9EECC4797417','Scott','Bridges','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
042507CCDD50','Colton','Kinney','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9F3BB81CC790','Brody','Goodman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A8D88E63A099','Simon','Park','M');
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4BD9B2621FE7','Addison','Medina','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35522FD65CAC','Charles','Sutton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4AA8F379FA2E','Vaughan','Lawson','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
69F75670D0C2','Felix','Haley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ason','Mclean','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
08E2134AFCB7','Brent','Faulkner','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
1BB0DEE71492','Kasimir','Fitzgerald','F');
_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
632F2CD4DC9E','Gregory','Berger','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EECF78E3635F','Damian','Key','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9EECC4797417','Scott','Bridges','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
042507CCDD50','Colton','Kinney','F');
IRST_NAME,LAST_NAME,GENDER) values
9F3BB81CC790','Brody','Goodman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A8D88E63A099','Simon','Park','M');
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4BD9B2621FE7','Addison','Medina','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35522FD65CAC','Charles','Sutton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4AA8F379FA2E','Vaughan','Lawson','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
69F75670D0C2','Felix','Haley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ason','Mclean','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
08E2134AFCB7','Brent','Faulkner','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) va
1BB0DEE71492','Kasimir','Fitzgerald','F');
_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
IRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
IRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
lues
25
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(15,'F812B9FD
Insert into CUSTOMER_DATA (
(16,'5F8F72F2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(17,'C54C5779
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(18,'D1EB489E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
(19,'50035120
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(20,'F35FB6B7
Insert into CUSTOMER_DATA (CUSTO
(21,'DBA752B4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(22,'23E5896B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(23,'480ED76A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(24,'FF7A36A3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(25,'E083DDBE
Insert into CUSTOMER_DATA (CUSTOMER
(26,'B7E4A6AD
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(27,'3BB44589
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(28,'D929DC13
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(30,'B8D4A245
Insert into CUSTOMER_DATA (CUSTOMER_ID,S
(31,'BFD9B89D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(32,'FC1E3D87
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(33,'60473650
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(34,'F44E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(35,'77F7D210
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCA
(36,'2DE41CE6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(37,'0A4B9BF9
25 | INTRODUCTION TO THE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(15,'F812B9FD
Insert into CUSTOMER_DATA (
(16,'5F8F72F2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(17,'C54C5779
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(18,'D1EB489E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
(19,'50035120
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(20,'F35FB6B7
Insert into CUSTOMER_DATA (CUSTO
(21,'DBA752B4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(22,'23E5896B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(23,'480ED76A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(24,'FF7A36A3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(25,'E083DDBE
Insert into CUSTOMER_DATA (CUSTOMER
(26,'B7E4A6AD
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(27,'3BB44589
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(28,'D929DC13
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(29,'B441636D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(30,'B8D4A245
Insert into CUSTOMER_DATA (CUSTOMER_ID,S
(31,'BFD9B89D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(32,'FC1E3D87
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(33,'60473650
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(34,'F44E0BF6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(35,'77F7D210
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCA
(36,'2DE41CE6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(37,'0A4B9BF9
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(15,'F812B9FD-0EC3-30FE-3D3D
Insert into CUSTOMER_DATA (
(16,'5F8F72F2-48DF-2507-6D34
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(17,'C54C5779-7BB8-4B02-D78E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(18,'D1EB489E-3FA6-C4CC-F923
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
(19,'50035120-B48A-4904-5498
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(20,'F35FB6B7-6B68-4E9E-5D1F
Insert into CUSTOMER_DATA (CUSTO
(21,'DBA752B4-5ABE-BEB5-360C
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(22,'23E5896B-9F6C-1E18-E671
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(23,'480ED76A-D8DC-3C45-1304
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(24,'FF7A36A3-C8B4-4305-9911
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(25,'E083DDBE-282B-2CCF-7D8A
Insert into CUSTOMER_DATA (CUSTOMER
(26,'B7E4A6AD-6331-09F1-43D8
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(27,'3BB44589-FD5E-8CAB-5E76
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(28,'D929DC13-0616-A3EC-0191
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9,'B441636D-CDE4-A840-E848
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(30,'B8D4A245-67F0-9386-7AAE
Insert into CUSTOMER_DATA (CUSTOMER_ID,S
(31,'BFD9B89D-08E3-6DB9-8F76
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(32,'FC1E3D87-3955-7F7C-F865
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(33,'60473650-8AB0-12F4-9C91
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0BF6-2F46-29D4-D9BC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(35,'77F7D210-C655-21DA-B8EC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCA
(36,'2DE41CE6-4298-109C-5598
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(37,'0A4B9BF9-BA50-5B21-7CFD
ORACLE STREAM EXPLORER
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3D3D-37ECB7A5B012','Zahir','Savage','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6D34-AA2E6FD663B8','Jin','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D78E-153879C9876A','Cha
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F923-CFB6005447F6','Chaim','Moses','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
5498-591F6F15A1FD','Lester','Fitzgerald','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5D1F-7EE852A11F8A','Otto','Dixon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
360C-572E52F02B4B','Reese','Gross','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E671-3BABD8E41E05','Griffit
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1304-EC7EECD18B25','Erasmus','Cobb','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9911-64F61FEC0CA7','Cooper','Barrett','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7D8A-0C1D77D93E5A','Stephen','Rivas','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
43D8-CA849A2FD796','Steven','Kramer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5E76-7640897ACCBE','Kirsten'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0191-8CE9B725F37F','Perry','Bender','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E848-B650635129C3','Alfonso','Velez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7AAE-11F0E30C582C','Brody','Craft','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,S
8F76-A9499A2B50D8','Trevor','Hinton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F865-EDD637D0558C','Nasim','Allis
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9C91-84B9E3B86DE7','Ali','Solomon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D9BC-4D35009C0D3B','Kennan','Schneider','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B8EC-869ED54F820D','Andrea','Santos','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCA
5598-4D74B8BEF432','Dillon','Cooke','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7CFD-82CBE0F7D9F7','Dante','Maxwell'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
37ECB7A5B012','Zahir','Savage','M');
CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AA2E6FD663B8','Jin','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
153879C9876A','Chadwick','Snyder','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CFB6005447F6','Chaim','Moses','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
591F6F15A1FD','Lester','Fitzgerald','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7EE852A11F8A','Otto','Dixon','M');
MER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
572E52F02B4B','Reese','Gross','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3BABD8E41E05','Griffit
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EC7EECD18B25','Erasmus','Cobb','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
64F61FEC0CA7','Cooper','Barrett','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0C1D77D93E5A','Stephen','Rivas','M');
_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CA849A2FD796','Steven','Kramer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7640897ACCBE','Kirsten'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8CE9B725F37F','Perry','Bender','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B650635129C3','Alfonso','Velez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11F0E30C582C','Brody','Craft','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9499A2B50D8','Trevor','Hinton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EDD637D0558C','Nasim','Allis
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
84B9E3B86DE7','Ali','Solomon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D35009C0D3B','Kennan','Schneider','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
869ED54F820D','Andrea','Santos','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D74B8BEF432','Dillon','Cooke','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
82CBE0F7D9F7','Dante','Maxwell'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
37ECB7A5B012','Zahir','Savage','M');
CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AA2E6FD663B8','Jin','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
dwick','Snyder','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CFB6005447F6','Chaim','Moses','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
591F6F15A1FD','Lester','Fitzgerald','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7EE852A11F8A','Otto','Dixon','M');
MER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
572E52F02B4B','Reese','Gross','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3BABD8E41E05','Griffith','Fields','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EC7EECD18B25','Erasmus','Cobb','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
64F61FEC0CA7','Cooper','Barrett','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0C1D77D93E5A','Stephen','Rivas','M');
_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CA849A2FD796','Steven','Kramer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7640897ACCBE','Kirsten','Malone','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8CE9B725F37F','Perry','Bender','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B650635129C3','Alfonso','Velez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11F0E30C582C','Brody','Craft','M');
CAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9499A2B50D8','Trevor','Hinton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EDD637D0558C','Nasim','Allison','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
84B9E3B86DE7','Ali','Solomon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D35009C0D3B','Kennan','Schneider','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
869ED54F820D','Andrea','Santos','F');
N_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D74B8BEF432','Dillon','Cooke','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
82CBE0F7D9F7','Dante','Maxwell','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
37ECB7A5B012','Zahir','Savage','M');
CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AA2E6FD663B8','Jin','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
dwick','Snyder','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CFB6005447F6','Chaim','Moses','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) valu
591F6F15A1FD','Lester','Fitzgerald','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7EE852A11F8A','Otto','Dixon','M');
MER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
572E52F02B4B','Reese','Gross','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
h','Fields','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EC7EECD18B25','Erasmus','Cobb','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
64F61FEC0CA7','Cooper','Barrett','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0C1D77D93E5A','Stephen','Rivas','M');
_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CA849A2FD796','Steven','Kramer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
,'Malone','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8CE9B725F37F','Perry','Bender','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B650635129C3','Alfonso','Velez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11F0E30C582C','Brody','Craft','M');
CAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9499A2B50D8','Trevor','Hinton','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
on','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
84B9E3B86DE7','Ali','Solomon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D35009C0D3B','Kennan','Schneider','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
869ED54F820D','Andrea','Santos','F');
N_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4D74B8BEF432','Dillon','Cooke','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
,'M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
MER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
N_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
es
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
MER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
N_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
26
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(38,'B71A0D5D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(39,'208CB1C1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(40,'8873F976
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(41,'3E4F0B65
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(42,'7DDBFAD6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(43,'1858CB09
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(44,'BF81BB5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(45,'CA2582D4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTR
(46,'0D333238
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(47,'656D6A68
Inse
(48,'7158556F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(49,'EA620BA5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(50,'DD5FF2F5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY
(51,'7DBF9ABB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(52,'F83662D2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(53,'4BA48031
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(54,'CBD8BCD1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(55,'617E92EB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,
(56,'935F0E85
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(57,'FB077D06
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(58,'E003CE4B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(59,'B645957B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(60,'A0617356
26 | INTRODUCTION TO THE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(38,'B71A0D5D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(39,'208CB1C1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(40,'8873F976
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(41,'3E4F0B65
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(42,'7DDBFAD6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(43,'1858CB09
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(44,'BF81BB52
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(45,'CA2582D4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTR
(46,'0D333238
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(47,'656D6A68
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(48,'7158556F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(49,'EA620BA5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(50,'DD5FF2F5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY
(51,'7DBF9ABB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(52,'F83662D2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(53,'4BA48031
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(54,'CBD8BCD1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(55,'617E92EB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,
(56,'935F0E85
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(57,'FB077D06
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(58,'E003CE4B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(59,'B645957B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(60,'A0617356
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(38,'B71A0D5D-544E-718B-EFE1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(39,'208CB1C1-5B8D-DB0E-C434
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(40,'8873F976-89FC-24A3-71B5
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(41,'3E4F0B65-7445-1CD8-8A31
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(42,'7DDBFAD6-D16B-F6A2-9F52
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(43,'1858CB09-1E4C-9272-1C10
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2-7111-B70F-78B8
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(45,'CA2582D4-40A0-C718-0F34
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTR
(46,'0D333238-D10C-8FF9-B2EB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(47,'656D6A68-7C70-02BC-32A8
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(48,'7158556F-CF03-05EA-D43F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(49,'EA620BA5-CD44-2487-87EA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(50,'DD5FF2F5-ADE0-18F6-1E86
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY
(51,'7DBF9ABB-4979-6BE6-2BBE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(52,'F83662D2-BC5A-1E00-4910
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(53,'4BA48031-B489-21BE-3EAD
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(54,'CBD8BCD1-0385-C7F0-6D94
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(55,'617E92EB-DDA0-3200-C13C
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,
(56,'935F0E85-9F11-3B2C-25B5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(57,'FB077D06-281F-C64C-99C4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(58,'E003CE4B-6AE8-046D-E0FB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(59,'B645957B-4C28-7F54-29FB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(60,'A0617356-3C2A-32B5-FC0D
ORACLE STREAM EXPLORER
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
EFE1-A9F171ECF738','Damian','Craig','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C434-071866F99570','Baker','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
71B5-A1AB6FE9D30E','Kelly','Briggs','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,
8A31-B5BAEED5CFA7','Eric','Rice','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9F52-6320A0B9C06F','Ashton','Mack','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1C10-F46F7D49BA6F','Maxwell','Mejia','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
78B8-F83A19F58E82','Marta','Winters','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0F34-12AD26AAEB3B','Harlan','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTR
B2EB-BCC1B1876767','Drew','Lynch','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
32A8-240E5B48332D','Felix','Cash','M');
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D43F-AF2B2DA28BEE','Clayton','Chambers','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
87EA-9CD352172BCC','Keaton','Woodward','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1E86-5E98C039D9B5','Ricardo','Ferreira','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY
2BBE-A782512D6AF5','Oliver','Hurst','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4910-53AE7FFB85C4','Channing','Hubbard','M'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3EAD-C2E245AEDC21','Peter','Frazier','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6D94-DAF8674087FA','Kieran','Farley','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C13C-CBAC5523526C','Monica','Atkins','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,
25B5-D0D8AC849427','Sharon','Wilkinson','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
99C4-46149A8555AB','Owen','Mccullough','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E0FB-9BEB6B991D33','Hayden','Wood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
29FB-7AD81627CB96','Mitsuko','Yakamoto','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FC0D-6170EE4E8D16','Garrison','Pugh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9F171ECF738','Damian','Craig','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
071866F99570','Baker','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A1AB6FE9D30E','Kelly','Briggs','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5BAEED5CFA7','Eric','Rice','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6320A0B9C06F','Ashton','Mack','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F46F7D49BA6F','Maxwell','Mejia','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F83A19F58E82','Marta','Winters','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
12AD26AAEB3B','Harlan','Wilcox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
BCC1B1876767','Drew','Lynch','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
240E5B48332D','Felix','Cash','M');
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AF2B2DA28BEE','Clayton','Chambers','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9CD352172BCC','Keaton','Woodward','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5E98C039D9B5','Ricardo','Ferreira','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY
A782512D6AF5','Oliver','Hurst','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
53AE7FFB85C4','Channing','Hubbard','M'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C2E245AEDC21','Peter','Frazier','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
DAF8674087FA','Kieran','Farley','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CBAC5523526C','Monica','Atkins','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,
D0D8AC849427','Sharon','Wilkinson','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
46149A8555AB','Owen','Mccullough','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9BEB6B991D33','Hayden','Wood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7AD81627CB96','Mitsuko','Yakamoto','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6170EE4E8D16','Garrison','Pugh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9F171ECF738','Damian','Craig','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
071866F99570','Baker','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A1AB6FE9D30E','Kelly','Briggs','F');
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5BAEED5CFA7','Eric','Rice','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6320A0B9C06F','Ashton','Mack','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F46F7D49BA6F','Maxwell','Mejia','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F83A19F58E82','Marta','Winters','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
12AD26AAEB3B','Harlan','Wilcox','M');
Y,FIRST_NAME,LAST_NAME,GENDER) values
BCC1B1876767','Drew','Lynch','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
240E5B48332D','Felix','Cash','M');
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AF2B2DA28BEE','Clayton','Chambers','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9CD352172BCC','Keaton','Woodward','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5E98C039D9B5','Ricardo','Ferreira','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A782512D6AF5','Oliver','Hurst','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
53AE7FFB85C4','Channing','Hubbard','M'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C2E245AEDC21','Peter','Frazier','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
DAF8674087FA','Kieran','Farley','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CBAC5523526C','Monica','Atkins','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D0D8AC849427','Sharon','Wilkinson','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
46149A8555AB','Owen','Mccullough','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9BEB6B991D33','Hayden','Wood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7AD81627CB96','Mitsuko','Yakamoto','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6170EE4E8D16','Garrison','Pugh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A9F171ECF738','Damian','Craig','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
071866F99570','Baker','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A1AB6FE9D30E','Kelly','Briggs','F');
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5BAEED5CFA7','Eric','Rice','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F46F7D49BA6F','Maxwell','Mejia','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F83A19F58E82','Marta','Winters','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
12AD26AAEB3B','Harlan','Wilcox','M');
Y,FIRST_NAME,LAST_NAME,GENDER) values
BCC1B1876767','Drew','Lynch','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
240E5B48332D','Felix','Cash','M');
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
AF2B2DA28BEE','Clayton','Chambers','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9CD352172BCC','Keaton','Woodward','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5E98C039D9B5','Ricardo','Ferreira','M');
,FIRST_NAME,LAST_NAME,GENDER) values
A782512D6AF5','Oliver','Hurst','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
53AE7FFB85C4','Channing','Hubbard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C2E245AEDC21','Peter','Frazier','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
DAF8674087FA','Kieran','Farley','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CBAC5523526C','Monica','Atkins','F');
FIRST_NAME,LAST_NAME,GENDER) values
D0D8AC849427','Sharon','Wilkinson','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
46149A8555AB','Owen','Mccullough','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9BEB6B991D33','Hayden','Wood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7AD81627CB96','Mitsuko','Yakamoto','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6170EE4E8D16','Garrison','Pugh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Y,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Y,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
rt into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
27
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENT
(61,'FC53DD7B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(62,'D4256448
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(63,'5AF682E5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(64,'C86B21AA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(65,'07AB86D5
Insert into CUSTOMER_DATA
(66,'20832A21
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(67,'EC4EFB21
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(68,'12A4FECC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(69,'BF2D39C3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(70,'13E4712C
Insert into CUSTOMER_DATA (CU
(71,'830373DE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(72,'0D562E79
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(73,'BE643FCC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
(74,'FD4483E9
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(75,'18555BC9
Insert into CUSTOMER_DAT
(76,'C32D18A2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(77,'EAB26AAA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(78,'375FC126
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
(79,'45BB2865
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(80,'B4D9B9DE
Insert into CUSTOMER_DATA
(81,'B1EBF0B6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(82,'5B69FE1B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(83,'FCFC70E9
27 | INTRODUCTION TO THE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENT
(61,'FC53DD7B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(62,'D4256448
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(63,'5AF682E5
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(64,'C86B21AA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(65,'07AB86D5
Insert into CUSTOMER_DATA
(66,'20832A21
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(67,'EC4EFB21
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(68,'12A4FECC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(69,'BF2D39C3
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(70,'13E4712C
Insert into CUSTOMER_DATA (CU
(71,'830373DE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(72,'0D562E79
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(73,'BE643FCC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
(74,'FD4483E9
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(75,'18555BC9
Insert into CUSTOMER_DAT
(76,'C32D18A2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(77,'EAB26AAA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(78,'375FC126
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
(79,'45BB2865
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(80,'B4D9B9DE
Insert into CUSTOMER_DATA
(81,'B1EBF0B6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(82,'5B69FE1B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(83,'FCFC70E9
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENT
(61,'FC53DD7B-393D-933D-B6B6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(62,'D4256448-B609-A58A-42B4
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(63,'5AF682E5-C189-E083-FE2D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(64,'C86B21AA-2789-A441-577A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(65,'07AB86D5-195F-85D0-2765
Insert into CUSTOMER_DATA
(66,'20832A21-D4E9-6549-973A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(67,'EC4EFB21-AFC6-A079-4FAE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(68,'12A4FECC-3BA8-1D1C-2B35
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(69,'BF2D39C3-6030-6938-EEEA
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(70,'13E4712C-463F-BAA7-8537
Insert into CUSTOMER_DATA (CU
(71,'830373DE-038B-95C5-467C
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(72,'0D562E79-16B8-B81D-6144
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(73,'BE643FCC-5F3B-E220-3670
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
(74,'FD4483E9-928B-91C9-BDB1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(75,'18555BC9-0B99-2F86-8FBB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(76,'C32D18A2-85AA-F3A3-5FD1
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(77,'EAB26AAA-DD16-0B56-4A6B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(78,'375FC126-B176-AC8D-D31D
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
(79,'45BB2865-CAA5-29E4-3E9B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(80,'B4D9B9DE-5912-CE87-235C
Insert into CUSTOMER_DATA
(81,'B1EBF0B6-F7C1-3DC4-3214
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(82,'5B69FE1B-2B9E-3919-F988
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(83,'FCFC70E9-6334-8B6C-D538
ORACLE STREAM EXPLORER
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENT
B6B6-0AEF570E14D3','Salvador','Holman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
42B4-B5AB0C7826FE','Tyrone','Phelps','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FE2D-919DAB0EAE96','Ryan','Molina','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
577A-0A9D586011A7','Paul','Mcintosh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2765-3FEEA7D2C666','Ralph','Perez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
973A-1FD322C7C710','Jennifer','Conelly','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4FAE-9B6BE9EA62A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2B35-BB478BDAD662','George','Morse','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
EEEA-4863600E7519','Matthew','Cole','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8537-A657BEB01D28','Blake','Benson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
467C-2573DFE51586','Debbie','Howell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6144-913CA61DA166','Na
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3670-FA99F7A1E507','Chandler','Dillon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
BDB1-A1A62FE2CA24','Hamilton','Rodriquez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8FBB-9ED4B5FBF1FC','Jeff','Mcdaniel','M');
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5FD1-8520D7892D40','Alfonso','Salazar','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4A6B-7BA332BCFB
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D31D-A589AA9B40EC','Maria','Nieves','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
3E9B-BE4010A3F9E2','Jonas','Sawyer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
235C-6CB165AD0FA0','Devin','Harper','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3214-35E7F5238C47','Aquila','Hatfield','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F988-10E4298CF4F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D538-7161AFA80739','Chaney','Pratt','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0AEF570E14D3','Salvador','Holman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5AB0C7826FE','Tyrone','Phelps','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
919DAB0EAE96','Ryan','Molina','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0A9D586011A7','Paul','Mcintosh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3FEEA7D2C666','Ralph','Perez','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1FD322C7C710','Jennifer','Conelly','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9B6BE9EA62AC','Raymond','Gould','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
BB478BDAD662','George','Morse','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
4863600E7519','Matthew','Cole','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A657BEB01D28','Blake','Benson','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2573DFE51586','Debbie','Howell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
913CA61DA166','Nataly','Shaffer','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FA99F7A1E507','Chandler','Dillon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
A1A62FE2CA24','Hamilton','Rodriquez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9ED4B5FBF1FC','Jeff','Mcdaniel','M');
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8520D7892D40','Alfonso','Salazar','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7BA332BCFB52','Calvin','Underwood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A589AA9B40EC','Maria','Nieves','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
BE4010A3F9E2','Jonas','Sawyer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6CB165AD0FA0','Devin','Harper','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35E7F5238C47','Aquila','Hatfield','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
10E4298CF4F4','Derek','Richard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7161AFA80739','Chaney','Pratt','M');
RY,FIRST_NAME,LAST_NAME,GENDER) values
0AEF570E14D3','Salvador','Holman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5AB0C7826FE','Tyrone','Phelps','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
919DAB0EAE96','Ryan','Molina','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0A9D586011A7','Paul','Mcintosh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3FEEA7D2C666','Ralph','Perez','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1FD322C7C710','Jennifer','Conelly','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C','Raymond','Gould','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
BB478BDAD662','George','Morse','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
4863600E7519','Matthew','Cole','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A657BEB01D28','Blake','Benson','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2573DFE51586','Debbie','Howell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
taly','Shaffer','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FA99F7A1E507','Chandler','Dillon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
A1A62FE2CA24','Hamilton','Rodriquez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9ED4B5FBF1FC','Jeff','Mcdaniel','M');
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8520D7892D40','Alfonso','Salazar','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
52','Calvin','Underwood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A589AA9B40EC','Maria','Nieves','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
BE4010A3F9E2','Jonas','Sawyer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6CB165AD0FA0','Devin','Harper','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35E7F5238C47','Aquila','Hatfield','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4','Derek','Richard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7161AFA80739','Chaney','Pratt','M');
RY,FIRST_NAME,LAST_NAME,GENDER) values
0AEF570E14D3','Salvador','Holman','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B5AB0C7826FE','Tyrone','Phelps','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
919DAB0EAE96','Ryan','Molina','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
0A9D586011A7','Paul','Mcintosh','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
3FEEA7D2C666','Ralph','Perez','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1FD322C7C710','Jennifer','Conelly','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
C','Raymond','Gould','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
BB478BDAD662','George','Morse','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
4863600E7519','Matthew','Cole','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A657BEB01D28','Blake','Benson','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2573DFE51586','Debbie','Howell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
taly','Shaffer','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
FA99F7A1E507','Chandler','Dillon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER)
A1A62FE2CA24','Hamilton','Rodriquez','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
9ED4B5FBF1FC','Jeff','Mcdaniel','M');
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8520D7892D40','Alfonso','Salazar','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
52','Calvin','Underwood','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A589AA9B40EC','Maria','Nieves','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GE
BE4010A3F9E2','Jonas','Sawyer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
6CB165AD0FA0','Devin','Harper','M');
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
35E7F5238C47','Aquila','Hatfield','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4','Derek','Richard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
7161AFA80739','Chaney','Pratt','M');
RY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
RY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
NDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
28
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(84,'471AF640
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(85,'332BF106
Insert into CUSTOMER_DATA (CU
(86,'6345109E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(87,'8B111703
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(88,'1C7FBB69
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(90,'BFBE734E
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(91,'B577A64A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(92,'FAFCE8D2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(93,'4ED62A82
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(94,'C89BD428
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(95,'1B5C67BF
Insert into CUSTOMER_DATA (CUSTOMER_
(96,'1829958B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(97,'ECA02944
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(98,'4E8DEBDD
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(100,'47CCF80B
--------------------------------
--
--------------------------------------------------------
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(1,'TSAMP','M','Good Mor
Samples?');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
28 | INTRODUCTION TO THE
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(84,'471AF640
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(85,'332BF106
Insert into CUSTOMER_DATA (CU
(86,'6345109E
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(87,'8B111703
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(88,'1C7FBB69
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(89,'95906C59
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(90,'BFBE734E
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(91,'B577A64A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(92,'FAFCE8D2
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(93,'4ED62A82
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(94,'C89BD428
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(95,'1B5C67BF
Insert into CUSTOMER_DATA (CUSTOMER_
(96,'1829958B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(97,'ECA02944
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(98,'4E8DEBDD
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(99,'0519C7FC
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(100,'47CCF80B
--------------------------------
-- Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(1,'TSAMP','M','Good Mor
Samples?');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
INTRODUCTION TO THE ORACLE STREAM EXPLOR
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
(84,'471AF640-42B5-2D28-F057
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(85,'332BF106-9B6B-C30C-1426
Insert into CUSTOMER_DATA (CU
(86,'6345109E-26FD-198B-F11F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(87,'8B111703-04FD-D5B6-1F9B
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(88,'1C7FBB69-266F-1F8D-902A
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
89,'95906C59-DA47-5700-55C9
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(90,'BFBE734E-C80D-35C5-B441
Insert into CUSTOMER_DATA (CUSTOMER_ID,
(91,'B577A64A-C1ED-18C4-01CF
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(92,'FAFCE8D2-1500-8712-38A6
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(93,'4ED62A82-FDCE-445C-F12F
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(94,'C89BD428-8B10-933B-CC58
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(95,'1B5C67BF-33F1-9F7B-D94A
Insert into CUSTOMER_DATA (CUSTOMER_
(96,'1829958B-0CC7-0B93-B804
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(97,'ECA02944-0A5F-023B-A292
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(98,'4E8DEBDD-2E3E-B59E-D785
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
99,'0519C7FC-3292-0B00-BB97
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(100,'47CCF80B-46A0-66C8-
--------------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(1,'TSAMP','M','Good Morning Mr. ',', would you be interested in trying out our
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ORACLE STREAM EXPLORER
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
F057-9F3CFE2072A2','Malik','Beard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1426-349909637024','Nathaniel','Cox','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F11F-410C93E52651','Elaine','Vargas','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
1F9B-A1C5A5077241','Ni
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
902A-4F987115CABC','Hu','Stein','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
55C9-8FBE929B09FD','Pamela','Hatfield','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B441-139B7296712E','Cyrus','Gay','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,
01CF-8F72EC65FA8C','Fletcher','Wiley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
38A6-710BDE770A4A','Fitzgerald'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
F12F-335745870917','Diane','Reese','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
CC58-5B01DDA827EA','Ingrid','Harrell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D94A-11560C15617C','Kenyon','Boyer','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
B804-8F023188DC2F','Emerson','Mcgowan','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A292-223016BA2B3C','Berk','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
D785-234767DC8522','Chancellor','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
BB97-E0464F7D64ED','Grant','Williamson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
-BA1A-2DA5FB862AEE','Judah','Salazar','M');
--------------------------------------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ning Mr. ',', would you be interested in trying out our
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
9F3CFE2072A2','Malik','Beard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
349909637024','Nathaniel','Cox','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
410C93E52651','Elaine','Vargas','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
A1C5A5077241','Nissim','Macias','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4F987115CABC','Hu','Stein','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8FBE929B09FD','Pamela','Hatfield','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
139B7296712E','Cyrus','Gay','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F72EC65FA8C','Fletcher','Wiley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
710BDE770A4A','Fitzgerald'
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
335745870917','Diane','Reese','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5B01DDA827EA','Ingrid','Harrell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11560C15617C','Kenyon','Boyer','M');
ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F023188DC2F','Emerson','Mcgowan','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
223016BA2B3C','Berk','
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
234767DC8522','Chancellor','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E0464F7D64ED','Grant','Williamson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2DA5FB862AEE','Judah','Salazar','M');
------------------------
Loading data into the table CUSTOMER_DATA
--------------------------------------------------------
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ning Mr. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
9F3CFE2072A2','Malik','Beard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
349909637024','Nathaniel','Cox','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
410C93E52651','Elaine','Vargas','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ssim','Macias','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4F987115CABC','Hu','Stein','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8FBE929B09FD','Pamela','Hatfield','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
139B7296712E','Cyrus','Gay','M');
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F72EC65FA8C','Fletcher','Wiley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
710BDE770A4A','Fitzgerald','Hughes','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
335745870917','Diane','Reese','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5B01DDA827EA','Ingrid','Harrell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11560C15617C','Kenyon','Boyer','M');
ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F023188DC2F','Emerson','Mcgowan','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
223016BA2B3C','Berk','Simon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
234767DC8522','Chancellor','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E0464F7D64ED','Grant','Williamson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2DA5FB862AEE','Judah','Salazar','M');
------------------------
--------------------------------------------------------
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ning Mr. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER
9F3CFE2072A2','Malik','Beard','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
349909637024','Nathaniel','Cox','M');
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
410C93E52651','Elaine','Vargas','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ssim','Macias','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
4F987115CABC','Hu','Stein','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8FBE929B09FD','Pamela','Hatfield','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
139B7296712E','Cyrus','Gay','M');
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F72EC65FA8C','Fletcher','Wiley','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
,'Hughes','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
335745870917','Diane','Reese','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
5B01DDA827EA','Ingrid','Harrell','F');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
11560C15617C','Kenyon','Boyer','M');
ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
8F023188DC2F','Emerson','Mcgowan','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Simon','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
234767DC8522','Chancellor','Sears','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
E0464F7D64ED','Grant','Williamson','M');
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
2DA5FB862AEE','Judah','Salazar','M');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ning Mr. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
ning Mr. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
STOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
Insert into CUSTOMER_DATA (CUSTOMER_ID,SCAN_ENTRY,FIRST_NAME,LAST_NAME,GENDER) values
29
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
Samp
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
Insert into C
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
Listing 1
For the implementation of the eye sc
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
the person's retina reading, the latitude and the longitu
eyeScan,latitude,longitude
C86B21AA
CB281A82
0A4B9BF9
E003CE4B
B645957B
B7C841D6
0028CF68
5AF682E5
C86B21AA
1B5C67BF
Listing 2
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
instead of the example shown in
URL:
Appendix B:
Considering that Oracle Stream Explorer allows sending the
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
manner. For instance, the case study
to persons walking around the store.
a greeting action?
Thanks to the
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
others,
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
29 | INTRODUCTION TO THE
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
Samples?');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
Insert into C
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
Listing 1. SQL script to create and populate the database tables.
For the implementation of the eye sc
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
the person's retina reading, the latitude and the longitu
eyeScan,latitude,longitude
C86B21AA-2789
CB281A82-EBF9
0A4B9BF9-BA50
E003CE4B-6AE8
B645957B-4C28
B7C841D6-DC7D
0028CF68-2264
5AF682E5-C189
C86B21AA-2789
1B5C67BF-33F1
Listing 2. Example of the CSV file used as eye scan stream.
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
instead of the example shown in
URL: http://www.ateam
Appendix B:
Considering that Oracle Stream Explorer allows sending the
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
manner. For instance, the case study
to persons walking around the store.
a greeting action?
Thanks to the Java Speech API
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
others, the first version of the Java Speech API
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
INTRODUCTION TO THE ORACLE STREAM EXPLOR
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
SQL script to create and populate the database tables.
For the implementation of the eye sc
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
the person's retina reading, the latitude and the longitu
eyeScan,latitude,longitude
2789-A441-577A-0A9D586011A7,30.831086,
EBF9-247F-8D4C-632F2CD4DC9E,30.425764,
BA50-5B21-7CFD-82CBE0F7D9F7,30.425764,
6AE8-046D-E0FB-9BEB6B991
4C28-7F54-29FB-7AD81627CB96,
DC7D-0C32-1402-58E0E06F0C74,30.674122,
2264-D591-C3F8-EECF78E3635F,30.674122,
C189-E083-FE2D-919DAB0EAE96,
2789-A441-577A-0A9D586011A7,
33F1-9F7B-D94A-11560C15617C,30.674122,
Example of the CSV file used as eye scan stream.
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
instead of the example shown in
http://www.ateam-oracle.com/wp
Appendix B: Creating a Message
Considering that Oracle Stream Explorer allows sending the
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
manner. For instance, the case study
to persons walking around the store.
a greeting action?
Java Speech API
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
the first version of the Java Speech API
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
ORACLE STREAM EXPLORER
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
Insert into CUSTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
USTOM_GREETING
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
SQL script to create and populate the database tables.
For the implementation of the eye scan stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
the person's retina reading, the latitude and the longitu
eyeScan,latitude,longitude
0A9D586011A7,30.831086,
632F2CD4DC9E,30.425764,
82CBE0F7D9F7,30.425764,
9BEB6B991D33,
7AD81627CB96,
58E0E06F0C74,30.674122,
EECF78E3635F,30.674122,
919DAB0EAE96,
0A9D586011A7,
11560C15617C,30.674122,
Example of the CSV file used as eye scan stream.
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
instead of the example shown in listing 2. The
oracle.com/wp-content/uploads/2015/02/EyeScanStream.csv
Creating a Message-
Considering that Oracle Stream Explorer allows sending the
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
manner. For instance, the case study implemented in this paper shown
to persons walking around the store. But what if
Java Speech API, create voice
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
the first version of the Java Speech API
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
SQL script to create and populate the database tables.
an stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
the person's retina reading, the latitude and the longitude from its location
0A9D586011A7,30.831086,-81.460571
632F2CD4DC9E,30.425764,-81.975556
82CBE0F7D9F7,30.425764,-81.975556
D33,-13.843414,-
7AD81627CB96,-13.843414,-
58E0E06F0C74,30.674122,-81.862946
EECF78E3635F,30.674122,-81.862946
919DAB0EAE96,-13.843414,-
0A9D586011A7,-13.843414,-
11560C15617C,30.674122,-81.862946
Example of the CSV file used as eye scan stream.
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
The complete version of the CSV file
content/uploads/2015/02/EyeScanStream.csv
-Driven Bean that Greets
Considering that Oracle Stream Explorer allows sending the output results from explorations to external systems, it
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
implemented in this paper shown
But what if the output results from the exploration were really transformed into
voice-based systems
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
the first version of the Java Speech API specification was released on October 26, 1998. From the technical
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(3,'WBACK','M','Hello Mr. ',', welcome back to the GAP.');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(4,'WBACK','F','Hello Mrs. ',', welcome back to the GAP.');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
an stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
de from its location.
81.460571
81.975556
81.975556
-55.371094
-55.371094
81.862946
81.862946
-55.371094
-55.371094
81.862946
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
complete version of the CSV file
content/uploads/2015/02/EyeScanStream.csv
Driven Bean that Greets
output results from explorations to external systems, it
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
implemented in this paper shown that it is p
the output results from the exploration were really transformed into
systems capable of spe
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
specification was released on October 26, 1998. From the technical
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
_MESSAGE) values
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(5,'BWELC','M','Hello Mr. ',', be welcome to our GAP Store!');
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(6,'BWELC','F','Hello Mrs. ',', be welcome to our GAP Store!');
an stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
complete version of the CSV file can be downloaded
content/uploads/2015/02/EyeScanStream.csv.
Driven Bean that Greets
output results from explorations to external systems, it
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
that it is possible to create custom greetings
the output results from the exploration were really transformed into
capable of speaking from text messages became
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
specification was released on October 26, 1998. From the technical
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third
(2,'TSAMP','F','Good Morning Mrs. ',', would you be interested in trying out our
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
_MESSAGE) values
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
(GREETING_ID,GREETING_CODE,GENDER,ICE_BREAK_MESSAGE,CUSTOM_MESSAGE) values
an stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
can be downloaded in the following
output results from explorations to external systems, it
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
ossible to create custom greetings
the output results from the exploration were really transformed into
aking from text messages became
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
specification was released on October 26, 1998. From the technical
point of view the Java Speech API is not part of the JDK implementation, it must be acquired from third-party speech
an stream, a CSV file with sample data need to be used. The listing 2 shows an
example of this file, containing only the first ten rows. The CSV file is structured having each row with three fields,
It is highly recommended that during the tests using Oracle Stream Explorer, the complete version of the file is used
the following
output results from explorations to external systems, it
would be interesting to leverage this capability to come up with solutions able to sense and respond in a contextual
ossible to create custom greetings
the output results from the exploration were really transformed into
aking from text messages became
possible. Originally developed by Sun Microsystems and in collaboration with companies like Apple, AT&T, IBM and
specification was released on October 26, 1998. From the technical
party speech
30
vendors
source
To demonstrate how the FreeTTS implementation can be used to
requests via JMS, listing
greeting using a human voice
package
import
import
import
import
import
import
import
import
import
import
@MessageDriven(activationConfig = {
public
30 | INTRODUCTION TO THE
vendors that provide their own implementations. One of the most popular
source speech synthesizer written entirely in Jav
To demonstrate how the FreeTTS implementation can be used to
requests via JMS, listing
greeting using a human voice
package com.oracle.fmw.ateam.fastdata;
import javax.annotation.PostConstruct;
import javax.annotation.PreDestroy;
import javax.ejb.ActivationConfigProperty;
import javax.ejb.EJBException;
import javax.ejb.MessageDriven;
import javax.jms.MapMessage;
import javax.jms.Message;
import javax.jms.MessageListener;
import com.sun.speech.freetts.Voice;
import com.sun.speech.
@MessageDriven(activationConfig = {
@ActivationConfigProperty(
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
@ActivationConfigProperty(
propertyName = "connectionFactoryJndiName
@ActivationConfigProperty(
propertyName = "destinationJndiName", prope
public class
private
@PostConstruct
private
}
@PreDestroy
private
INTRODUCTION TO THE ORACLE STREAM EXPLOR
that provide their own implementations. One of the most popular
speech synthesizer written entirely in Jav
To demonstrate how the FreeTTS implementation can be used to
requests via JMS, listing 3 shows an example of a MDB that leverages the Java Speech API to synthesize the
greeting using a human voice.
com.oracle.fmw.ateam.fastdata;
javax.annotation.PostConstruct;
javax.annotation.PreDestroy;
javax.ejb.ActivationConfigProperty;
javax.ejb.EJBException;
javax.ejb.MessageDriven;
javax.jms.MapMessage;
javax.jms.Message;
javax.jms.MessageListener;
com.sun.speech.freetts.Voice;
com.sun.speech.freetts.VoiceManager;
@MessageDriven(activationConfig = {
@ActivationConfigProperty(
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
@ActivationConfigProperty(
propertyName = "connectionFactoryJndiName
@ActivationConfigProperty(
propertyName = "destinationJndiName", prope
GreetingListener
private Voice voice;
@PostConstruct
private void allocateVoice() {
VoiceManager voiceManager = VoiceManager.getInstance();
voice = voiceManager.getVoice("kevin16");
voice.allocate();
@PreDestroy
private void deallocateVoice() {
if (voice !=
ORACLE STREAM EXPLORER
that provide their own implementations. One of the most popular
speech synthesizer written entirely in Jav
To demonstrate how the FreeTTS implementation can be used to
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
com.oracle.fmw.ateam.fastdata;
javax.annotation.PostConstruct;
javax.annotation.PreDestroy;
javax.ejb.ActivationConfigProperty;
javax.ejb.EJBException;
javax.ejb.MessageDriven;
javax.jms.MapMessage;
javax.jms.Message;
javax.jms.MessageListener;
com.sun.speech.freetts.Voice;
freetts.VoiceManager;
@MessageDriven(activationConfig = {
@ActivationConfigProperty(
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
@ActivationConfigProperty(
propertyName = "connectionFactoryJndiName
@ActivationConfigProperty(
propertyName = "destinationJndiName", prope
GreetingListener implements
Voice voice;
allocateVoice() {
VoiceManager voiceManager = VoiceManager.getInstance();
voice = voiceManager.getVoice("kevin16");
voice.allocate();
deallocateVoice() {
(voice != null) {
that provide their own implementations. One of the most popular
speech synthesizer written entirely in Java.
To demonstrate how the FreeTTS implementation can be used to
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
com.oracle.fmw.ateam.fastdata;
javax.annotation.PostConstruct;
javax.ejb.ActivationConfigProperty;
freetts.VoiceManager;
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
propertyName = "connectionFactoryJndiName
propertyName = "destinationJndiName", prope
implements MessageListener {
allocateVoice() {
VoiceManager voiceManager = VoiceManager.getInstance();
voice = voiceManager.getVoice("kevin16");
deallocateVoice() {
that provide their own implementations. One of the most popular implementations is the
To demonstrate how the FreeTTS implementation can be used to create a voice
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
propertyName = "connectionFactoryJndiName", propertyValue = "jms/connFact
propertyName = "destinationJndiName", propertyValue = "jms/greetingQueue")
MessageListener {
VoiceManager voiceManager = VoiceManager.getInstance();
voice = voiceManager.getVoice("kevin16");
implementations is the
a voice-based system that listen gre
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
", propertyValue = "jms/connFact
rtyValue = "jms/greetingQueue")
VoiceManager voiceManager = VoiceManager.getInstance();
implementations is the FreeTTS, an open
based system that listen gre
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
propertyName = "destinationType", propertyValue = "javax.jms.Queue"),
", propertyValue = "jms/connFact"),
rtyValue = "jms/greetingQueue")})
n open-
based system that listen greeting
3 shows an example of a MDB that leverages the Java Speech API to synthesize the
"),
})
31
}
Listing 3
In the
javax.jms.MapMessage
results when the JMS
generated by the exploration, as shown in figure 24.
31 | INTRODUCTION TO THE
}
@Override
public
}
}
Listing 3. Implementation of the MDB using the Java
In the onMessage
javax.jms.MapMessage
results when the JMS
generated by the exploration, as shown in figure 24.
INTRODUCTION TO THE ORACLE STREAM EXPLOR
voice.deal
}
@Override
public void onMessage(Message message) {
if (message
MapMessage mapMessage =
String iceBreakMessage =
String customerName =
String customMessage =
String greeting =
try
}
}
}
Implementation of the MDB using the Java
onMessage()method implementation shown in listing 3, the received message is transformed into a
javax.jms.MapMessage, because this is the type of message that Oracle Stream Explorer sends the output
results when the JMS-based target is used. Also, the attributes
generated by the exploration, as shown in figure 24.
ORACLE STREAM EXPLORER
voice.deallocate();
onMessage(Message message) {
(message instanceof
MapMessage mapMessage =
String iceBreakMessage =
String customerName =
String customMessage =
String greeting =
try {
mapMessage = (MapMessage) message;
iceBreakMessage = mapMessage.getString("iceBreakMessage");
customerName = mapMessage.getString("customerName");
customMessage = mapMessage.getString("custom
greeting = iceBreakMessage + customerName + customMessage;
voice.speak(greeting);
catch (Exception ex) {
throw new
Implementation of the MDB using the Java
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
based target is used. Also, the attributes
generated by the exploration, as shown in figure 24.
locate();
onMessage(Message message) {
instanceof MapMessage) {
MapMessage mapMessage = null
String iceBreakMessage = null
String customerName = null;
String customMessage = null
String greeting = null;
mapMessage = (MapMessage) message;
iceBreakMessage = mapMessage.getString("iceBreakMessage");
customerName = mapMessage.getString("customerName");
customMessage = mapMessage.getString("custom
greeting = iceBreakMessage + customerName + customMessage;
voice.speak(greeting);
(Exception ex) {
new EJBException(ex);
Implementation of the MDB using the Java Speech API.
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
based target is used. Also, the attributes
generated by the exploration, as shown in figure 24.
MapMessage) {
null;
null;
;
null;
mapMessage = (MapMessage) message;
iceBreakMessage = mapMessage.getString("iceBreakMessage");
customerName = mapMessage.getString("customerName");
customMessage = mapMessage.getString("custom
greeting = iceBreakMessage + customerName + customMessage;
voice.speak(greeting);
EJBException(ex);
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
based target is used. Also, the attributes obtained from the message matches with the ones
mapMessage = (MapMessage) message;
iceBreakMessage = mapMessage.getString("iceBreakMessage");
customerName = mapMessage.getString("customerName");
customMessage = mapMessage.getString("custom
greeting = iceBreakMessage + customerName + customMessage;
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
from the message matches with the ones
iceBreakMessage = mapMessage.getString("iceBreakMessage");
customerName = mapMessage.getString("customerName");
customMessage = mapMessage.getString("customMessage");
greeting = iceBreakMessage + customerName + customMessage;
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
from the message matches with the ones
iceBreakMessage = mapMessage.getString("iceBreakMessage");
greeting = iceBreakMessage + customerName + customMessage;
method implementation shown in listing 3, the received message is transformed into a
, because this is the type of message that Oracle Stream Explorer sends the output
from the message matches with the ones
32
From the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
FreeTTS implementation also n
EAR) or
deployment are:
32 | INTRODUCTION TO THE
From the compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
FreeTTS implementation also n
EAR) or installed
deployment are:
» $FREETTS/lib/cmulex.jar
» $FREETTS/lib/
» $FREETTS/lib/en_us.jar
» $FREETTS/lib/freetts.jar
INTRODUCTION TO THE ORACLE STREAM EXPLOR
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
FreeTTS implementation also n
installed directly into the
$FREETTS/lib/cmulex.jar
$FREETTS/lib/cmu_us_kal.jar
$FREETTS/lib/en_us.jar
$FREETTS/lib/freetts.jar
ORACLE STREAM EXPLORER
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
FreeTTS implementation also need to be deployed, either toge
into the Java EE application server
cmu_us_kal.jar
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
eed to be deployed, either toge
application server classpath. The libra
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
eed to be deployed, either together with the MDB (In case of packaging it using an
classpath. The libra
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
ther with the MDB (In case of packaging it using an
classpath. The libraries that need to be considered for
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
ther with the MDB (In case of packaging it using an
ries that need to be considered for
compilation perspective, the only library from the FreeTTS implementation that needs to be available in the
classpath is the $FREETTS/lib/freetts.jar library. But from the deployment perspective, some other libraries from the
ther with the MDB (In case of packaging it using an
ries that need to be considered for
CONN E C T W I T H U S
CONN E C T W I T H U S
blogs.oracle.com/oracle
facebook.com/oracle
twitter.com/oracle
oracle.com
CONN E C T W I T H U S
blogs.oracle.com/oracle
facebook.com/oracle
twitter.com/oracle
Oracle Corporation, World Headquarters
500 Oracle Parkway
Redwood Shores, CA 94065, USA
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,contents hereof are subject towarranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantfitness for a particulformed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by ameans, elec Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective Intel and Intel Xeon are trademarks or regiare trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo aretrademarks or registered trademarks of Adv Introduction to the Oracle Stream ExplorerMarchAuthor: Ricardo FerreiraReviewers:
Oracle Corporation, World Headquarters
500 Oracle Parkway
Redwood Shores, CA 94065, USA
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,contents hereof are subject towarranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantfitness for a particular purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by ameans, electronic or mechanical, for any purpose, without our prior written permission.
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective
Intel and Intel Xeon are trademarks or regiare trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo aretrademarks or registered trademarks of Adv
Introduction to the Oracle Stream ExplorerMarch 2015 Author: Ricardo Ferreira Reviewers: Peter Farkas, Prabhu Thukkaram
Oracle Corporation, World Headquarters
Redwood Shores, CA 94065, USA
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,contents hereof are subject to change without notice. This document is not warranted to be errorwarranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchant
ar purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by a
tronic or mechanical, for any purpose, without our prior written permission.
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective
Intel and Intel Xeon are trademarks or registered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo aretrademarks or registered trademarks of Advanced Micro Devices. UNIX is a register
Introduction to the Oracle Stream Explorer
, Prabhu Thukkaram
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,change without notice. This document is not warranted to be error
warranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantar purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are
formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by atronic or mechanical, for any purpose, without our prior written permission.
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective
stered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo are
anced Micro Devices. UNIX is a register
Worldwide Inquiries
Phone: +1.650.506.7000
Fax: +1.650.506.7200
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,change without notice. This document is not warranted to be error
warranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantar purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are
formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by atronic or mechanical, for any purpose, without our prior written permission.
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective
stered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo are
anced Micro Devices. UNIX is a registered trademark
Worldwide Inquiries
Phone: +1.650.506.7000
+1.650.506.7200
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only,change without notice. This document is not warranted to be error-free, nor subject to any other
warranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantar purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are
formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by a
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective
stered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo are
ed trademark of The Open Group.0115
Copyright © 2014, Oracle and/or its affiliates. All rights reserved. This document is provided for information purposes only, and the free, nor subject to any other
warranties or conditions, whether expressed orally or implied in law, including implied warranties and conditions of merchantability or ar purpose. We specifically disclaim any liability with respect to this document, and no contractual obligations are
formed either directly or indirectly by this document. This document may not be reproduced or transmitted in any form or by any
Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners.
stered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo are
of The Open Group.0115