Staff upgrade bib fuzzy match (MRC)

Owned by Corinna Baksik
Last updated: Aug 14, 2024



REMEMBER: Only designated staff members are permitted to run jobs. Refer to the matrix at Jobs & Batch Processes to see which staff members are approved to run which jobs.

Usage

This import profile is designed for records downloaded from OCLC.

This profile will merge/overlay the existing Alma record and should be used with caution. The merge routine used is the same one that is used for OCLC WorldCat Updates (see the "Prefer OCLC" column). 

Due to the fuzziness of the matching, this profile should be used with great caution to upgrade bibs that lack OCLC numbers. 

Requirements

  • Vendor records must be in UTF8 encoding format. Non-Latin fields should be in 880 pairs. 
  • Input file should be in MARC binary format (not MARCXML)
  • Vendor record must have either:
    • 035 $a(OCoLC)... If it does not match to an existing Alma record it will not load. 
    • OR
    • Both 001 with OCLC number and 003 with (OCoLC). Alma will construct the appropriate 035 based on these values.

(warning)  Notice: Presence of both 035 $a(OCoLC) and the 001/003 (OCoLC) combination will bring in an extraneous 035 field with the $a content, because Alma creates a new 035 based on the 001/003. This import profile uses a special norm rule to remove the 003 prior to the load, so that only the existing 035 from OCLC will be created. It is assumed that the recordd being imported with this profile come from OCLC and have the 035 already. If this is not the case this profile should not be used. Contact LTS with any questions. 

Matching

This profile uses the "Multi-step match method" which will perform a sequence of tests. If a bib fails to match on the first test, Alma moves on to the second test, then third. If there is still no match, the record is not loaded. 


More details on the Alma routines listed below may be found on the ExL site. 

TestAlma routine nameSummary
1Unique OCLC Identifier Match Method

Must match on OCLC number in 035 $a or $z

If there is a match, testing is done. 

2

ISBN (exact subfield match)/ 024 / 035 Match Method

(or ISSN version of serials)

  1. Incoming 020 $a must match Alma 020 $a, or
  2. Incoming 020 $z must match Alma 020 $z, or
  3. 024 or 035 match

If there is a match, testing is done. 

3

Title Statement Extended Fuzzy Non-Serial Match Method

(or Serial version)

Alma attempts to find records with at least one of the following matching IDs:
  • ISSN
  • ISBN
  • LCCN
  • CODEN
  • OCLC unique number
  • Other system number (035 field)

AND (???)  if there is a match on one of the above identifiers, year and format must match as follows: 

  • Publication is one year or less as determined by:
    • MARC 21 positions 7-10 in the 008 field
  • Identical format as determined by the record type specified in the LDR


IF THERE IS STILL NO MATCH:

If the incoming record has authors, Alma will test for author+title match as follows: 

  • MARC 21 Title: 245 abknp 
  • MARC 21 Author: 100 a-d,jq,u; 110 a-e,n,u; 111 a,c-e,n,q,u; 700 a-d,jq,u; 710 a-e,I,n,u; 711 a,c-e,I,j,n,q,u
  • AND
  • LDR format match AND pub year within 1


IF THERE IS STILL NO MATCH:

Alma attempts to find records that have at least one matching title with:

  • MARC 21 Title: 245 abknp 
  • AND
  • LDR format match AND pub year within 1

In the event that multiple matches are found for any step, Alma will choose the bib that was added most recently 

Results

  • No inventory will be created. 
  • If a record is locked by a staff member it will not be overlaid.
  • The incoming record will upgrade the Alma record, with certain fields being protected. See list of protected fields here: https://harvardwiki.atlassian.net/wiki/x/6aGVAgE

Example use case

  • You have a set of Alma records that lack OCLC numbers and you wish to upgrade them with copy from OCLC.Â