pingest.pl should use the rdeleted parameter when ingesting record attributes

[Jason Stephenson]

Evergreen version 3.4.1 (affects version 3.3.0+)
PostgreSQL version 9.6.16
OpenSRF Version 3.2.0 (N/A)

The pingest.pl program takes an option to specify which record attributes to reingest when run without the --skip-attrs option specified. However, it does not specify a value for the rdeleted parameter of the metabib.reingest_record_attributes function. The rdeleted parameter defaults to a TRUE value, which bypasses the pattr_list parameter. Thus, the value(s) of the --attr parameter(s) are ignored and all record attributes are reingested by pingest.pl.

This can be fixed by using the deleted column value of the record being ingested.

[Jason Stephenson (jstephenson)]

Here's a branch that uses the deleted field of the biblio.record_entry being ingested when metabib.reingest_record_attributes is run in the pingest.pl support script:

https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/dyrcona/lp1862652-pingest-reingest-record-attrs-rdeleted

[Jane Sandberg (sandbergja)]

Thanks, Jason. To check my understanding: the deleted field of the biblio.record_entry being ingested will always be false, right? Since pingest.pl doesn't (yet?) have a way to revive deleted records into undeleted ones?

[Jason Stephenson (jstephenson)]

Jane, not in all cases. You can send a list of record ids into pingest.pl. That list can include deleted records.

Basically, what I found was that all of the fields were being reingested on non-deleted records, when I asked for only certain fields to be reingested. This patch corrects that for non-deleted records and still ingests all fields when a deleted record is being reingested.

Hope that makes sense.

[Jason Stephenson (jstephenson)]

BTW, this patch makes a huge performance difference when doing an attribure only ingest, particularly if a list of attributes is specified.

I'm presently ingesting only the mr_hold_format attribute of our non-deleted records, and it is going much more quickly after applying this patch.

[Jane Sandberg (sandbergja)]

A belated thank you for the answer to my question, Jason. This works well -- I saw the same performance improvement that Jason described. Pushed to master, rel_3_8, and rel_3_7.