In Spring 2018 I began cleaning up a large collection of e-resource records from a single vendor. Some records were coming from our knowledgebase, and some from our ILS (which in turn may have been sourced from worldcat or the provider’s website).
I began by determining the number and quality of the records coming from the knowledgebase. Of 11,931 records, 7,539 were brief (determined by the code ssib), usually indicating nothing more than title, author and provider. In comparison, 41,250 records were present in our ILS, of varying quality, with 83,353 of generally good quality available from the vendor.
I reported that I thought the best strategy for handling records from this vendor would be through the ILS, with records coming from the vendor site. We continued to subscribe to the collections in the knowledgebase, but without the MARC feed for records, to facilitate any future moves to an LSP.
I then created a spreadsheet which tracked the number of records by collection, the catkeys from the current ILS (which we would need to delete the old records), the last time the records were updated in the ILS and updated by the vendor on their site. From here I assigned each collection a priority level, based on the disparity between current records and number of records available from the provider, the quality of records, and if the collection had been updated recently (which was especially important for non-static collections).
I downloaded records from the vendor, which required me to unzip and join the files in MarcEdit, and create task lists based on common problems (such as excessive 653 fields, and $h[electronic resource] being present in the 245 field).The records were cleaned up, database IDs were added to the 856 field, existing catkeys were compiled into .csv files to be deleted, and documentation and instructions were prepared for uploading the files on a bi-yearly basis.
In total, over 65,000 new records were added to the catalogue, 20,000 were removed, and 20,000 were preserved.
The whole reconciliation process was documented in OneNote and Confluence, where we keep track of all of the action required for all of our e-Resources. I also set a reminder to check the provider’s site quarterly for updated records.