Here's my slow, simple script for comparing the titles of the bibs our holdings were attached to in Ex Libris Alma against the bibs in our legacy Evergreen library system.
Necessary because of missing accents in many of the Alma bibs (thanks anglophone partners!) and outright mismatches based on OCLC numbers (often data quality issues in our bibs).
SRU and a simple REST API #FTW although it's taken all week to process 600K records.
@dbs The implication here is that you can't even import data into your Ex Libris instance.
Is that normal? That doesn't sound normal.
@victoria We were joining an existing consortium of 14 other libraries, with a shared catalogue of bibliographic records (bibs). In theory it will save us all work in the long run.
But 13 of the other 14 libraries are unilingual anglophones, so their bibs got in first, and after that if there's a match based on OCLC number you use their bib. No quality comparisons: just first one in wins.
@victoria My assumption, based on early returns, is that most of the problems are with older records. These date back when we were converting from a card catalogue to an electronic system. The anglophones wouldn't have bothered to have systems that could handle accents.
@dbs first one in wins? Bad.
The assumption about old systems that didn't support accents makes sense, but that just further highlights the need to collaboratively amend records to improve accuracy.
If only there was a way to do that with software.
code4lib.social is a GLAM-themed Mastodon Instance.