Closed cbeer closed 7 months ago
Seems like: Poppy added field for timestamp on instance reflecting whether holdings/items have changed. We need to assess if that is in fact true, and then see if we can use the Poppy info to save some indexing work.
May also require assessing impact on indexing memory issues.
Tests completed
Based on the caveats this does not fix #980 but it does allow us to clean up some code.
Couple of things that came up with exploring the data:
table | Number of rows per table | How to access instance |
---|---|---|
sul_mod_inventory_storage.instance | 10,284,710 | n/a |
sul_mod_inventory_storage.holdings_record | 12,366,458 | instanceid field |
sul_mod_inventory_storage.items | 11,429,559 | holdingsrecordid field |
One odd thing about the holdings_record is there doesn't seem to be any empty instanceid fields (assuming I queried correctly). Some holdings have same instanceids (see WHERE instanceid = '6b68aea8-2230-52a9-b008-ce8f21309f90')
Possibly fixes https://github.com/sul-dlss/searchworks_traject_indexer/issues/980