Date post: | 01-Jan-2016 |
Category: |
Documents |
Upload: | benjamin-buck |
View: | 26 times |
Download: | 0 times |
TopX 2.0TopX 2.0——
A (Very) Fast Object-Store A (Very) Fast Object-Store for for
Top-k XPath Query Top-k XPath Query ProcessingProcessingMartin Theobald
Stanford University
Ralf SchenkelMax-Planck Institute
Mohammed AbuJarourHasso-Plattner Institute
“Native XML data base systems can store schemaless data ... ”
“Data management systems control data acquisition, storage, and retrieval. Systems evolved from flat files … ”
“XML-QL: A Query Language for XML.”
“Native XML Data Bases.”
“Proc. Query Languages Workshop, W3C,1998.”
“XML queries with an expres- sive power similar to that of Datalog …”
sec
article
sec
par
bib
par
title “Current Approaches to XML Data Manage-ment”
itempar
title inproc
title
//article[.//bib[about(.//item, “W3C”)] ]//sec[about(.//, “XML retrieval”)] //par[about(.//, “native XML databases”)]
“What does XML add for retrieval? It adds formal ways …”
“w3c.org/xml”
sec
article
sec
par “Sophisticated technologies developed by smart people.”
par
title “The
XML Files”
par
title “TheOntology Game”
title“TheDirty LittleSecret”
bib
“There, I've said it - the "O" word. If anyone is thinking along ontology lines, I would like to break some old news …”
title
item
url“XML”
RANKINGRANKINGRANKINGRANKING
VAGUENESSVAGUENESSVAGUENESSVAGUENESS
EARLY PRUNINGEARLY PRUNINGEARLY PRUNINGEARLY PRUNING
From the INEX ’03-’05 IEEE Collection
Ontology/Large Thesaurus
WordNet,OpenCyc, etc.
Ontology/Large Thesaurus
WordNet,OpenCyc, etc.
SASA
Relational DBMS BackendUnified Text & XML Schema
Relational DBMS BackendUnified Text & XML Schema
Random Access
Top-kQueueTop-kQueue
Scan Threads
Scan Threads
CandidateQueue
CandidateQueue
Indexer/Crawler Indexer/Crawler
Frontends• Web Interface • Web Service • API
Frontends• Web Interface • Web Service • API
• Selectivities• Histograms• Correlations
• Selectivities• Histograms• Correlations
Index MetadataIndex Metadata
TopX 1.0 Query Processor
TopX 1.0 Query Processor
Sequential Access
SASA SASA
• Path Conditions• Phrases & Proximity• Other Full-Text Op’s
• Path Conditions• Phrases & Proximity• Other Full-Text Op’s
Expensive PredicatesExpensive Predicates
RARA
Probabilistic Candidate
Pruning
Probabilistic Candidate
Pruning
Probabilistic Index AccessScheduling
Probabilistic Index AccessScheduling
Dynamic Query
Expansion
Dynamic Query
Expansion
Non-conjunctiveTop-k XPath
Query Processing
Non-conjunctiveTop-k XPath
Query Processing
RARA
JDBCJDBC
2.0
Data Model
XML trees (no XLink/ID/IDRef) Pre-/postorder ranges for the structural index Redundant full-content text nodes
<article>
<title>XML Data Management </title> <abs>XML management systems vary widely in their expressive power. </abs> <sec> <title>Native XML Data Bases. </title> <par>Native XML data base systems can store schemaless data. </par> </sec></article>
“xml data manage xml manage system vary wide expressive power native xml data base native xml data base system store schemaless data“
“native xml data base native xml data base system store schemaless data“
“xml data manage”
articlearticle
titletitle absabs secsec
“xml manage system vary wide
expressivepower“
“native xml data base”
“native xml data base system store schemaless data“
titletitle parpar
1 6
2 1 3 2 4 5
5 3 6 4
“xml data manage xml manage system vary
wide expressive power native xml native
xml data base system store schemaless data“
ftf (“xml”, article1 ) = 4ftf (“xml”, article1 ) = 4
ftf (“xml”, sec4 ) = 2ftf (“xml”, sec4 ) = 2
“native xml data base native xml data base system store schemaless data“
Scoring Model [INEX ‘05/’06/’07]
XML-specific variant of Okapi BM25 (originating from probabilistic IR on unstructured text)
Content Index (Tag-Term Pairs) Element Freq. Element Statistics
bib[“transactions”]vs.
par[“transactions”]
bib[“transactions”]vs.
par[“transactions”]
TopX 1.0: Relational Schema Precompute & materialize scoring model into combined inverted index over tag-term pairs Supports sorted access (by MaxScore) and random access (by DocID)
sec[“xml”]
Select DocID, Pre, Post, Score From TagTermIndex Where tag=‘sec’ and term=‘xml’ Order by MaxScore desc, DocID desc
Pre asc, Post Desc
SASA
Select Pre, Post, Score From TagTermIndex Where DocID=3 and tag=‘sec’ and term=‘xml’ Order by Pre Asc, Post Desc
RARA
Two B+trees
Top-k XPath on a Relational Schema [VLDB ’05]
• Content-only (CO) & “structure enriched” queries: //sec[about(.//, “XML”) and about(.//title, “native”]//par[about(.//,
“retrieval”)]
Sequentially (mostly) scan each index list in desc. order of MaxScore Hash-join element blocks by DocID in-memory Do “some” incremental XPath evaluation using Pre/Post indices Aggregate Score along connected path fragments Use variant of Fagin’s threshold algorithm for top-k-style early termination
sec[“xml”] title[“native”] par[“retrieval”]
article
RARA
Expensive predicate probes (RA) to the structure index (3rd B+tree)
Non-conjunctive XPath evaluations Dynamically relax content- & structure-related query conditions
(top-k results entirely driven by score aggregations for content & structure cond.’s)
• Content-and-structure (CAS) queries: //article//sec[about(.//, “XML”)]
Select Pre, Post From TagIndex Where DocID=2123 and Tag=‘article’Order by Pre asc, Post desc
sec[“xml”]
SASA
Top-k XPath on a Relational Schema [VLDB ’05]
1.0
Relational Schema (cont’d)
20,810,942 distinct tag-term pairs for 4.38 GB Wikipedia collection
20,810,942 distinct tag-term pairs for 4.38 GB Wikipedia collection
sec[“xml”] article
No shredding into DTD-specific relational schema! No DTD at all for INEX Wikipedia!
1,107 distinct tags1,107 distinct tags
Relational Schema (cont’d)
2-dimensional source of redundancy Full-content scoring model (#terms times avg. depth of a text node 6.7 for INEX Wiki) De-normalized relational schema
High overhead in the architecture (Java->JDBC->DBMS & back) Element-block sizes are data-driven, not easy to control layout on disk Hashing too slow compared to very efficient in-memory merge-joins
Content Index Structure Index
(4+4+4+4+4+4+4) bytes X 567,262,445 tag-term pairs
16 GB
(4+4+4+4) bytes X 52,561,559 tags
0.85 GB
TopX 2.0: Object-Oriented Storage
2 15 0.92DocID
10 8 0.5
23 48 0.8
45 87 0.2
MaxSore
1MaxSore
DocID
sec[“xml”]
0
title[“xml”]
122,564
…
par[“xml”]
432,534
(4+4+4+4+4+4+4) X 567,262,445
Relational: 16 GB
4 X 456,466,649+ (4+4+4) X 567,262,445
Object-oriented: 8.6 GB
(+ (4+4) X 20,810,942 = 166 MB for the offset index)
B
17
3B
L
L
Binary file
B – Element block separatorL – Index list separator
Group element blocks with similar MaxScore into document blocks of fixed length (e.g. 256KB)
Sort element blocks within each document block by DocID
Supports Sorted access by MaxScore Merge-joins by DocID
Raw disk access
Object-Oriented Storage w/Block-Mergingsec[“xml”]
0
title[“xml”]
122,564L
B…BB
2
1
B
5B
…
…B…
BB
3
6B
Doc
umen
t Blo
ck
MaxSore
MaxSore
Merging Document Blocks
Sequential access and efficient merge-joins on top of large document blocks
sec[“xml”]
B…
BB
2
1
B
5B
…B…
BB
3
6B
…
par[“retrieval”]
B…BB
5
2
B
7B
B…BB
6
9B
//sec[about(.//, “XML”)] //par[about(.//, “retrieval”)]
SASA
1.0
0.8
0.8
0.7
Compressed Number Encoding Multi-attribute (4), double-nested block-index structure
Delta encoding only works for DocID (and to some extent for Pre) No specific assumptions on distributions of Pre/Post or Score
No Unary or Huffman coding (prefix-free but additional coding table)
Sophisticated compression schemes may be expensive to decode No Zip, etc.
But known number ranges DocID [1, 659,388] -> 3 bytes (2543 = 16,387,064, lossless) Pre/Post [1, 43,114] -> 2 bytes (2542 = 64,516, lossless) Score [0,1] -> rounded to 1 byte (254 buckets, lossy)
Variable-length byte encoding w/leading length-indicator byte
5 bytes
10 bytes
Some more tricks… Dump leading histogram blocks into index list headers
Histograms only for index lists that exceed one document block (<5% of all lists) Own native compare methods for DocID, Pre/Post Decode only Score for arithmetic op’s
( Mostly perform pointer operations at qp time)
Incrementally read & process precomputed memory image for fast top-k queries on top of large disk blocks
36 b
ytes
10
sec[“xml”]
score
freq
DB1 (256 KB)
…
DB2 (256 KB) DBl (256 KB)
… … …1.0
0.9
0.8
0.8
1.0
0.9
0.9
0.2
1.0
0.9
0.7
0.6
SA Scheduling Look-ahead Δi through precomputed
score histograms Knapsack-based optimization of
Score Reduction
RA Scheduling 2-phase probing:
Schedule RAs “late & last”
i.e., cleanup the queue if
Extended probabilistic cost model for integrated SA & RA scheduling
Block Access Scheduling [VLDB ’06]
Inverted Block-Index(256KB doc-blocks)
Δ3,3 = 0.2Δ3,3 = 0.2Δ1,3 = 0.8Δ1,3 = 0.8
SA
SASA
SA SA
SA
RARA
Object Storage Summary
• 567,262,445 tag-term pairs• 20,810,942 distinct tag-term pairs• 20,815,884 document blocks (256KB)• 456,466,649 element blocks
• 4,703,385,686 total bytes (8.3 bytes/tag-term pair)
• 52,561,559 tags (elements)• 1,107 distinct tags• 2,323 document blocks (256KB)• 8,999,193 element blocks
• 246,601,752 total bytes (4.7 bytes/tag)
4.38 GB Wikipedia XML sources
Structure IndexCont
ent I
ndex
(incl
. his
togr
ams)
Conclusions & Outlook
Scalable and efficient XML-IR with vague search Mature system, reference engine for INEX topic development &
interactive tracks [VLDB Special Issue on DB&IR Integration ‘08]
Brand-new TopX 2.0 prototype Very efficient reimplementation in C++ Object-oriented XML storage, moderate compression rates 10—20 times better sequential throughput than relational
More features Generalized proximity search, graph top-k Updates (gaps within document blocks) XQuery Full-Text (top-k-style bounds over IF, For-Let) …