Tuesday, January 31, 2012
Announcement to Hold Bibliographic edits on Aleph bibs prior to extraction
The message below went out to Cataloger-L on Dec. 15. Bernardo's single file extraction had to be redone in portions and rerun on Thursday night so no data edited in existing bibs on Dec. 14 was lost.
Going forward, any new bibliographic records added to Euclid/Aleph after Dec. 15 may be edited as they will go out when we can resume our weekly AEX service (after loading new authorities and corrected bibs). Any edits to those bibliographic records added to the catalog after Dec. 15 and edited after this date will be preserved and not overwritten. Only the corrected headings will be affected. LTI will return the bibs in the same state as before they were shipped, except for headings corrections based on new authorities matching.
Please continue to hold edits to bibliographic records that were in Euclid before the extraction on Dec. 15. You may edit records in OCLC and save edited records in your save file until after LTI corrected bibs are loaded, if you wish). Please validate headings in OCLC prior to saving so the headings are authorized and you don't overlay corrected bibliographic headings.
Note that 1.2 million bibliographic records in Emory's bibliographic file had headings corrections in this bib refresh process. That amounts to 1/3 of Emory'd database. This means there is a 1 out of 3 chance any edits you make in those database records extracted on Dec. 16, 2011 AFTER that date may be overwritten if the record you edited had unauthorized headings. This does not apply to Aleph alphabetic fields in the bib such as LKR, TKR, etc. and does not apply to edits in HOL and ITEM records.
Note that 1.2 million bibliographic records in Emory's bibliographic file had headings corrections in this bib refresh process. That amounts to 1/3 of Emory'd database. This means there is a 1 out of 3 chance any edits you make in those database records extracted on Dec. 16, 2011 AFTER that date may be overwritten if the record you edited had unauthorized headings. This does not apply to Aleph alphabetic fields in the bib such as LKR, TKR, etc. and does not apply to edits in HOL and ITEM records.
Update: Dec. 2011
There has been a delay in the production load schedule. DPS wanted to do the entire process in test and did not have enough time to run all the loads in test prior to MLK holiday weekend as scheduled.
Test loads were competed on Jan. 30. There will be one week review and followup on issues with ExLibris. When issues are resolved, DPS can schedule the loads. When a schedule is known, DPS will communicate with all libraries' management and a general message will go out regarding downtime, interim operations at service desks, cataloging stoppages, etc.
Status of LTI Loads in test
Monday, Jan. 30 2012
========
I met with AMS authority team to initiate our review of Aleph authority loads and corrected bibs. We will review sample record keys supplied by DPS and report anomalies and issues to ExLibris support.
========
jeff
please feel free to test records.
the authority load and re-indexing services take almost 6 hours.
the BIB load and re-indexing services take 35 hours.
bernardo
-----------------------------------------
Bernardo Gomez
Library IT
Emory University
404-727-4948
LTI followup to subject string validation question and unlinked headings report
The "unlinked headings" report lists headings which were not linked (at $a level) to an LC or an LTI authority record.
Should you be interested, an optional report of "partially validated headings" is available on request. This report lists headings that were linked at $a, but there were subsequent subfields in the heading of which one or more could not be validated by an authority record, or from one of out tables (floats, geographic, language, etc.) In the "names" file, the bulk of these are name/title headings where the name links; in the subjects, our tables allow us to validate many of the "main + subdivision" strings, so there are fewer reported.
Should you be interested, an optional report of "partially validated headings" is available on request. This report lists headings that were linked at $a, but there were subsequent subfields in the heading of which one or more could not be validated by an authority record, or from one of out tables (floats, geographic, language, etc.) In the "names" file, the bulk of these are name/title headings where the name links; in the subjects, our tables allow us to validate many of the "main + subdivision" strings, so there are fewer reported.
Impact of Emory's provisional record policy for Sirsi and LTI project
I looked at the file "EMORY-AUP114.LCS" and found 304 authority records beginning:
"World War, 1914-1918" and 839 beginning "World War, 1939-1945".
However, there are only 3 authority records in the entire LC authority file which are subdivided by "United States" (e.g. "World War, 1914-1918|zUnited States..."). You received all 3. The situation with "World War, 1939-1945, United States" is similar, with only 6 LC authority records, all of which were provided for Emory. All of the above LC authorities contain the 667 note: "|aRecord generated for validation purposes."
Though LC does not, in general, create authority records for "Authorized heading + float", a project begun several years ago did a pass through the bib file, creating authority records for full heading strings, including floats or subdivisions (e.g. geographic). The above 667 field was added to each of these machine-generated authority records.
In the absence of a national-level authority record, some local systems create authority records for the full subject heading string. LTI's "provisional" option did this for the library's former Sirsi system, creating (hundreds of?) thousands of records with an "ltp" number in the 001. Sorry, but I can't tell how many provisional records would have been created. I can say that, in addition to "validated" heading strings (main + subdivision), the provisional records included all the unlinked headings, since they were based on "no national level authority record". There were over 300,000 entries on the "unlinked headings" report generated by the AUP.
FYI, as of 1/23, the numbers for the entire LC authority file were:
names: 8,316,971 -- lcsh: 406,712 -- kids: 897
"World War, 1914-1918" and 839 beginning "World War, 1939-1945".
However, there are only 3 authority records in the entire LC authority file which are subdivided by "United States" (e.g. "World War, 1914-1918|zUnited States..."). You received all 3. The situation with "World War, 1939-1945, United States" is similar, with only 6 LC authority records, all of which were provided for Emory. All of the above LC authorities contain the 667 note: "|aRecord generated for validation purposes."
Though LC does not, in general, create authority records for "Authorized heading + float", a project begun several years ago did a pass through the bib file, creating authority records for full heading strings, including floats or subdivisions (e.g. geographic). The above 667 field was added to each of these machine-generated authority records.
In the absence of a national-level authority record, some local systems create authority records for the full subject heading string. LTI's "provisional" option did this for the library's former Sirsi system, creating (hundreds of?) thousands of records with an "ltp" number in the 001. Sorry, but I can't tell how many provisional records would have been created. I can say that, in addition to "validated" heading strings (main + subdivision), the provisional records included all the unlinked headings, since they were based on "no national level authority record". There were over 300,000 entries on the "unlinked headings" report generated by the AUP.
FYI, as of 1/23, the numbers for the entire LC authority file were:
names: 8,316,971 -- lcsh: 406,712 -- kids: 897
Update on Extraction from Sirsi of Emory Local Notes in Authority Records
Thursday Jan. 5, 2012
This extrcact serves for backup to search for missing 690 information.
Marian was thinking of a full extract of authority files containing 690s for use in a potential project to merge 690 notes only into matching 001/010 after we do the full restore of corrected bibs, new authorities, and re-index.
On 1/5/12 12:37 PM, "Kelley, Marian" <libmhk@emory.edu> wrote:
>The authority record 690 note file was output in a pipe-delimited file
>that has three columns:
>001 | 010 | 690 |
>
>There was no output of complete records.
>
>Marian Kelley
LTI Summary of Base Bib File Refresh
Monday, Jan. 9 2012.
1,206,609 changed records in: EMORY-AUP114.CHG. This is the number of corrected bibs (new authorized headings) from LTI process.
The output files and reports from the 4th Quarter 2011
>are on the LTI FTP server for you to retrieve
>using your AEX/AUP user name and password:
>
>EMORY-AUP114.LSB -- 85,252,816 bytes -- text listing of bib heading
>changes
>EMORY-AUP114.LSA -- 11,139,799 bytes -- text listing of auth rec changes
>EMORY-AUP114.ULH -- 12,845,350 bytes -- text listing of unlinked headings
>EMORY-AUP114.CHG -- 1,528,203,240 bytes -- changed bib records
>EMORY-AUP114.LCN -- 859,023,778 bytes -- new/revised LC name auth recs
>EMORY-AUP114.LCS -- 84,582,331 bytes -- new/revised LC subj auth recs
>EMORY-AUP114.LCG -- 382,607 bytes -- new/revised LC Genre/Form auth recs
>EMORY-AUP114.LCM -- 12,823,225 bytes -- new/revised NLM MeSH auth recs
>EMORY-AUP114.LCDEL -- 0 bytes -- delete authority records
>EMORY-AUP114-nfichg.ACR -- 463,314 bytes -- list of non-filing ind.
>changes
Subscribe to:
Posts (Atom)