EPO Worldwide Patent Statistical Database James ROLLINSONEuropean Patent OfficePatent InformationPost Grant Renewal Fees

3rd Annual Patent analysis workshop on "The Output of R&D activities: Harnessing the Power of Patents Data" - IPTS Seville, 13-14 June, 2011

the story

• Early prototype developed by OECD (Science, Technology, Industry Section)

• EPO took over responsibility for production• First distributed in April 2006• Restricted distribution in development period• Since October 2007 publicly available• It has established itself as a database of choice for

– statisticians– academics– policy advisors

EPO Worldwide Patent Statistical Database (PATSTAT)

For the first time, the EPO has provided an off-line worldwide patent database:

– we provide the data (tables) & database model– we show how to store the data– researchers can work with the entire database on

a standard laptop PC

Thanks to the OECD (Science, Technology, Industry Section) in Paris, France, for their earlier database work

PATSTAT Central Database

Application

Applicants

Inventors

Classes

PublicationsCitations

Priorities

Families

Physical tables in the April 2011 editionTable name Number of rows per table in

April 2011

TLS201_APPLN 68.481.300

TLS202_APPLN_TITLE 49.502.431

TLS203_APPLN_ABSTR 18.746.129

TLS204_APPLN_PRIOR 29.706.641

TLS205_TECH_REL 21.301.338

TLS206_PERSON 38.418.130

TLS207_PERS_APPLN 138.684.682

TLS208_DOC_STD_NMS 17.275.375

TLS209_APPLN_IPC 164.335.118

TLS210_APPLN_N_CLS 46.788.300

TLS211_PAT_PUBLN 76.668.550

TLS212_CITATION 104.598.819

TLS214_NPL_PUBLN 15.898.451

TLS215_CITN_CATEG 18.675.863

TLS216_APPLN_CONTN 1.850.333

TLS217_APPLN_ECLA 105.467.978

TLS218_DOCDB_FAM 60.312.073

TLS219_INPADOC_FAM 68.481.300

Application_ID is the central key

Relationships between applications

• Priorities under the Paris Convention (INID 30) Table 204

• References to other related domestic Table 216patent documents (INID 60)

• addition• continuation in part • division• reissue• substitute

• Data related to other International Table 201Conventions (INID 80) eg. PCT

• Technical relations Table 205

Persons: applicants (assignees, grantees, proprietors) and inventors

• Person is physical or legal entity• Person can be applicant AND inventor• doc_std_name: max. 30 char.• Names and addresses from most recent publication• "docdba" elements are used (as received by EPO)• USPTO data: 1/3 of sequence data missing person_name

and doc_std_name_id might not match• special txt files for US data with individual elements

Persons and address data

• person_ctry_code from data format "docdb" (standardised), except US, EP

• person_ctry_code: coverage is 50%, not JP• Special EP cases: "data withheld", "the designation of the

inventor has not yet been filed"• EP data reflect the last changes (from Bulletin database)

Authority EP Other (GB, IE,...)Publication type Published applications Published patents All AllRange Jan. 1976- Nov. 2005 DOCDB XML OECD patent database ESPACE BULLETIN DOCDB XMLRange Nov. 2005-today (*) USPTO website USPTO website ESPACE BULLETIN DOCDB XML

(*) Sept. 2005 to today for published applications

US

Patent publications

• For publications having more than one occurrence (example EP or WO A9), only the last occurrence is loaded

• Invalid or empty dates: 9999-12-31• Abstracts and titles are attributed to the application, in case of

multiple occurrences:– English has priority– the most recent is selected

• publn_first_grant:

Citations

• Reliable data for: AP, AU, BE, CZ, DE, EP, ES FR, NL, SG, US, WO

• Batches only for: JP, DK, LU, GR, TR• Euro-PCT: npl_biblio = "See references of WO 0046271A1"• Patent citations hidden in NPL citations:

– npl_publn_id > 0– npl_citn_seq_nr > 0 , and– cited_pat_publn_id > 0

TLS209: IPC-8 classification

• IPC 1-7 symbols are NOT given• 570.000 documents have no IPC8 but a IPC1-7 symbol

– 10% published after 2006!– 10% have an ECLA symbol

• IPC8 classes are aggregated and de-duplicated at application and simple family level

• advanced symbols are given• Jan. 2011: revision of IPC-8 -

core symbols are no longer maintained

TLS217: ECLA classification

• Nanotech codes: – epo_class_scheme = ICO– epo_class_symbol = Y01N6:00

• Environmentally Sound Technologies (in September 2010)– epo_class_scheme = ICO– epo_class_symbol = Y02B10:00, Y02C10:00,...

• Schemes covered:– EC– ICO– ECNO– IDT

DOCDB Simple Patent Familyversus

INPADOC Extended Family

• INPADOC Extended Family – is covering a technology– might be slight differences in technical content– members do not have to share more than one priority with at

least one other member, directly or indirectly

• DOCDB Simple Patent Family– is covering one invention – technical content covered is identical– members have to share identical priority pictures

DOCDB Simple versus INPADOC Extended

• INPADOC Extended Family – broadest definition of a patent family– supports identification technological trends– supports definition of geographical coverage

• DOCDB Simple Patent Family– subset of INPADOC Extended Family– particularly suited for prior art search– tailored to the needs of EPO examiners

DOCDB Simple versus INPADOC Extended

one INPADOC extended =four DOCDB simple families

first filing & division of

continuation in part

continuation in part

continuation in part

Shoe with anatomical protection

Patent families: TLS218 and TLS219

• TLS218: simple family– Family-identifier is created in DOCDB– Family-identifier is a surrogate key, unique but change is

not excluded (application might change family)– Surrogate keys remain the same through PATSTAT

editions

• TLS219 INPADOC family– Family-identifier is created in PATSTAT– inpadoc_family_id changes with every edition of PATSTAT

Changes in September 2010 edition

• SEA ==> 0 - citations introduced during search• APP ==> 1 - citations introduced by the applicant• EXA ==> 2 - citations introduced during examination• OPP ==> 3 - citations introduced during opposition• 115 ==> 4 - citations introduced according to Art 115 EPC

• ISR ==> 5 - citations from the International Search Report• SUP ==> 6 - citations from the Supplementary Search Report• CH2 ==> 7 - citations introduced during the Chapter 2 phase of the

PCT

Changes in September 2010 edition• change of source to DOCDB XML for element

PUBLN_FIRST_GRANT.• The table TLS211_PAT_PUBLN contains the column

PUBLN_FIRST_GRANT. If this has the value '1' , then that publication is the 'first grant'.

• In April 2010, the method for calculating this was based on the publication kind code representing a grant in each country, and then selecting the earliest publication.

• In September 2010 we use the 'public-availability' tag in the DOCDB XML product from the EPO.

Changes in September 2010 edition• New table: Table TLS221_INPADOC_PRS containing INPADOC

worldwide legal status data was created and integrated into the PATSTAT database structure.

• However it was produced on a test basis only

• to be available as of April 2011 edition but will have to be acquired separately.

Changes in April 2011 edition• Table TLS201_APPLN: New permanent unique application identifier

introduced in APPLN_ID.

• With the April 2011 edition, the DOCDB "doc-id" unique and stable identifier has been used to populate APPLN_ID instead of creating a PATSTAT-edition-specific surrogate key

• (but not for the artificial applications in PATSTAT).

• DOCDB attribute "doc-id" contains a stable and unique identifier that will allow for linking up a number of EPO raw data products through the application in a reliable way.

• This attribute will remain the same across PATSTAT editions and will always refer to the same combination of application authority, application number and application kind.

TLS201_APPLN

• publications• "True" applications

if appln_ID < 900 000 000• Unpublished priorities if appln_ID

between 900,000,001 and 906,479,936• "dummy" D2 applications from • citations if appl_ID between 907.000.001 and

908.692.290• appl_kind = W for PCT applications• PCT origin is given in internat_appln_ID• ipr_type: UM, DP, PI

Changes in April 2011 edition

• Table TLS209_APPLN_IPC: IPC Core Level symbols are no longer maintained in WIPO ST8.

• Until September 2011, both Advanced and Core sets of symbols, now Adv

• The IPC Core symbols eliminated from DOCDB, unless a publication had a Core symbol but no Advanced symbol, or when two families are joined (by priority changes)

• Take care with re-using old SQL queries with IPC

Changes in April 2011 edition

• Table TLS221_INPADOC_PRS containing INPADOC worldwide legal status

• PRS INPADOC Worldwide Legal Status is a separate raw data product

• http://www.epo.org/searching/subscription/raw.html• Pricing is 910 euros for the full database , annually• In PATSTAT compatible CSV format• Subscription 2011 (on physical carrier) EUR 1,090 • If customers wish to purchase only a one-off copy,

then the current backfile price for PRS 14.11 is applied, that is EUR 910.

Planned developments at last Seville meeting

• EST to become part of ECLA/ICO codes (Y02) • Extension with worldwide legal data: PRS in CSV

format with PATSTAT "ApplicationID"• Unique application ID stable appl_ID• publn_first_grant from routine to DOCDB XML feed• Number of claims for US data• PCT address data• Standardisation of names (external input)• PATSTAT visualisation project

Planned developments at last Seville meeting

•EST to become part of ECLA (Y02)•Extension with worldwide legal data: PRS in CSV format with PATSTAT “application id”•Unique application id stable appln_id•Publn_first_grant from routine to DOCDB_XML feed•PATSTAT visualisation project

•Number of claims for US data•PCT address data•Standardisation of names (external input)

Planned developments today

1. Number of claims for US & EP data2. More address data – PCT, FR, GB, ES3. Standardisation of names (external input)4. Cited filing applications5. Add indication of ISA for WO citations 6. Identify Non Patent Literature NPL differently7. Japanese & US classification schemes8. Japanese patent abstracts in english9. Remove JP & US classifications from TLS210

Number of claims

1. US data: relates to granted patents only– (A documents until 2000, B1 or B2 documents

afterwards)

2. EP data: relates to – published applications from 1978– granted patents from 2006.

3. Add new column PUBLN_CLAIMS to TLS211_pat_publn

More address data – PCT, FR, GB, ES

• Subscribers have asked for more address data

• French, British , Spanish patent offices have also shown interest

• WIPO have kindly provided PCT data in special file

• Needs to be loaded into DocDB EPO database first

• May be in an XML ‘blob’

Cited filing applications

TLS212_CITATIONPAT_PUBLN_IDCITN_IDCITED_PAT_PUBLN_IDNPL_PUBLN_IDPAT_CITN_SEQ_NRNPL_CITN_SEQ_NRCITN_ORIGIN

CITED_APPLN_ID

Change in procedure allows unpublished applications

Add indication of ISA for WO citations

• Country code identifying the patent authority performing the International Search Report

• The new column CITN_GENER_AUTH will not be populated for other citations, only PCT ( ISA ) ones.

• International Search Authority

Add indication of ISA for WO citations

TLS212_CITATIONPAT_PUBLN_IDCITN_IDCITED_PAT_PUBLN_IDNPL_PUBLN_IDPAT_CITN_SEQ_NRNPL_CITN_SEQ_NRCITN_ORIGIN

CITED_APPLN_IDCITED_GENER_AUTH

Identify Non Patent Literature NPL differently

• Instead of creating a new surrogate key, extract the 9 digit XP number from DOCDB and use this 9 digit number as surrogate key, removing the leading zeros

• These numbers are not allocated sequentially.

• As NPL_PUBLN_ID is the unique primary key, the duplicate cited NPL texts are NOW REMOVED in PATSTAT.

• A side effect of this change is that the table TLS214 will be reduced (fewer rows).

Japanese & US classification schemes

• Both the Japanese and the US patent offices have agreed that their classification schemes - FI/Fterm and DOCUS respectively - may now be exchanged.

• From August 2011 onwards, we will cover the following classification schemes : EC, ICO, ECNO, IDT, FI, Fterm and DOCUS.

• Additional tables to hold the new classification- schemes :

• TLS222_APPLN_JP_CLASS

• TLS223_APPLN_DOCUS

Japanese patent abstracts in english

• The Japanese patent office has now agreed that the english language abstracts of japanese patents may be exchanged.

• An additional 9 million abstracts in English for JP publications will be added.

• Question to existing users: can the load files be increased in size?

Remove JP & US classifications from TLS210

• There will be major new tables in PATSTAT for US and for JP classifications

• Leaving the US and JP symbols in tls210_appln_n_cls may cause PATSTAT subscribers to make errors in their research.

• Remove all JP and US symbols from table tls210_appln_n_cls

Questions?

...you have the floor!

Thank you for your attention

James ROLLINSONpatstat@epo.org

1616--17 November 2011 Washington 17 November 2011 Washington