[
https://issues.apache.org/jira/browse/SOLR-15191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17308347#comment-17308347
]
David Smiley commented on SOLR-15191:
-------------------------------------
Apparently the test infrastructure randomly picks not only {{EnumFieldType}}
but also {{EnumField}}. The {{assumeFalse}} only checks for one of them. I
suggest checking for neither, and thus merely demand that there is docValues.
I'm not sure if this test could work with other scenarios but it was designed
specifically for EnumFieldType. Makes sense?
[~thomas.woeckinger] why did you pick {{method : enum}} for the JSON facet
algorithm? If I recall (my memory is hazy on this), we're going to use the
hash algorithm so why specify enum? Just to show that it picks one that can
actually work based on the field's config?
> Faceting on EnumFieldType does not work if allBuckets, numBuckets or missing
> is set
> -----------------------------------------------------------------------------------
>
> Key: SOLR-15191
> URL: https://issues.apache.org/jira/browse/SOLR-15191
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: Facet Module, FacetComponent, faceting, search,
> streaming expressions
> Affects Versions: 8.7, 8.8, 8.8.1
> Reporter: Thomas Wöckinger
> Assignee: David Smiley
> Priority: Major
> Labels: easy-fix, pull-request-available
> Fix For: 8.9
>
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> Due to Solr-14514 FacetFieldProcessorByEnumTermsStream is not used if
> allBuckets, numBuckets or missing parma is true.
> As fallback FacetFieldProcessorByHashDV is used which
> FacetRangeProcessor.getNumericCalc(sf) on the field. EnumFileType is not
> handled currently, so a SolrException is thrown with BAD_REQUEST and
> 'Expected numeric field type'
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
