[
https://issues.apache.org/jira/browse/SOLR-14477?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris M. Hostetter updated SOLR-14477:
--------------------------------------
Attachment: SOLR-14477.patch
Status: Open (was: Open)
patch updated with TestCloudJSONFacetSKG inspired by work being done in
SOLR-13132...
* beef up index to include single valued fields that also get randomized facet
testing
* refactor test's (randomizable) {{TermFacet}} data struct to be a little more
flexible, making it easier to add new options
* add randomization of "prefix" option
* for good measure, also add in randomization of:
** perSeg
** prelim_sort
So far this all seems to be beasting well ... but because TestCloudJSONFacetSKG
works by firing off "verification" queries to check the results for each
bucket, the test has some limitations in the indexes and option combos it
creates – so before calling this issue "done" i also want to incorporate some
of the other test work in progress in SOLR-13132 to randomize comparisons of
the results for the same request when only the facet "method" option is varied,
to give myself a little more piece of mind that this fix is complete.
> relatedness() values can be wrong when using 'prefix'
> -----------------------------------------------------
>
> Key: SOLR-14477
> URL: https://issues.apache.org/jira/browse/SOLR-14477
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Chris M. Hostetter
> Assignee: Chris M. Hostetter
> Priority: Major
> Attachments: SOLR-14477.patch, SOLR-14477.patch
>
>
> Another {{relatedness()}} bug found in json facet's while working on
> increased test coverage for SOLR-13132.
> if the {{prefix}} option is used when doing a terms facet, then the
> {{relatedess()}} calculations can be wrong in some situations -- most notably
> when using {{limit:-1}} but i'm pretty sure the bug also impacts the code
> paths where the (first) {{sort}} (or {{prelim_sort}} is computed against the
> {{relatedness()}} values.
> Real world impacts of this bug should be relatively low since i can't really
> think of any practical usecases for using {{relatedness()}} in conjunction
> with {{prefix}}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
