Add Works Error - Too many creators?

[NoraRamsey]

When adding works on the UI:
We used a doi (10.48550/arXiv.2211.05100) to autofill information in the "descriptions" tab. There were over 50 creators in the file and when we tried to save and submit got an error message. Graham and I assume it's because the record has too many creators, but it's beneficial to increase any creator limit if that exists.

Error message: "We're sorry, but something went wrong: If you are the application owner check the logs for more information."

[cziaarm]

error: Document contains at least one immense term in field=\"suggest\" (whose UTF8 encoding is longer than the max length 32766), all of which were skipped. Please correct the analyzer to not produce such terms
The prefix of the first immense term is: '[91, 123, 34, 99, 114, 101, 97, 116, 111, 114, 95, 103, 105, 118, 101, 110, 95, 110, 97, 109, 101, 34, 58, 34, 98, 105, 103, 115, 99, 105]...',
original message: bytes can be at most 32766 in length; got 46942. Perhaps the document has an indexed string field (solr.StrField) which is too large

392 creators means creators_tesim is someway north of the 32766 byte limit, but surely creators_tesim is a solr.Textfield not a solr.StrField... I am confused adding needs discussion

[cziaarm]

@jeremyf @orangewolf : All my solr sleuthing has suggested that creators is already a solr.TextField

I'm going on the fact that it is a _tesim field and *_tesim is in schema.xml as a solr.TextField

It seems unlikely that any other fields would cause the error... creator_search is the only other candidate, but that is under the bytes threshold in this case and is also a _tesim

I may be misunderstanding something fundamental. I hope so as other options for this issue seem at best very awkward.

[j-basford]

'needs discussion' label - is that with BL or with SoftServ? Just want to check if there is more info that might be needed from our end

[cziaarm]

This one #442 is the ticket that best describes the problem.

Possibly quite a general hyrax issue when indexing fields with lots and lots of values (here's at least one other person who has bumped into it, a while back samvera-deprecated/sufia#3007).

Either there's a really neat solr config trick that will make this go away otherwise it might require some deep thought about how the creator field is structured, handled and indexed.

It is a tricky one TBH

¯\_(ツ)_/¯