[jira] [Commented] (LUCENE-10070) "count all" faceting functionality counts deleted docs for multiple implementations

["Greg Miller (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-10070?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17404811#comment-17404811 ] 

Greg Miller commented on LUCENE-10070:
--------------------------------------

See tests in this PR for repro: https://github.com/apache/lucene/pull/263

> "count all" faceting functionality counts deleted docs for multiple implementations
> -----------------------------------------------------------------------------------
>
>                 Key: LUCENE-10070
>                 URL: https://issues.apache.org/jira/browse/LUCENE-10070
>             Project: Lucene - Core
>          Issue Type: Bug
>          Components: modules/facet
>            Reporter: Greg Miller
>            Priority: Minor
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> A few different {{Facets}} implementations supporting a "count all" style constructor that allows the user to not pass in a {{FacetsCollector}} instance. It advertises that it's equivalent to using a {{FacetsCollector}} populated with a {{MatchAllDocsQuery}}, but more efficient. It looks like, with the exception of {{FastTaxonomyFacetCounts}}, none of the implementations correctly account for deleted documents (have a look at {{FastTaxonomyFacetCounts}} for a correct example that consults "live docs."
> From what I can tell, the affected implementations are:
>  * SortedSetDocValueFacetCounts
>  * ConcurrentSortedSetDocValueFacetCounts
>  * LongValueFacetCounts
>  * StringValueFacetCounts
> I'll attach a PR shortly illustrating unit tests I wrote that confirm the bug.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org