[
https://issues.apache.org/jira/browse/SOLR-12516?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hoss Man reassigned SOLR-12516:
-------------------------------
Assignee: Hoss Man
Attachment: SOLR-12516.patch
attaching basically the same patch as yesterday -- just updated to account for
a few conflicts & test-tweaks needed because of the SOLR-12326 changes.
> JSON "range" facets can incorrectly refine subfacets for buckets
> ----------------------------------------------------------------
>
> Key: SOLR-12516
> URL: https://issues.apache.org/jira/browse/SOLR-12516
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: Facet Module
> Reporter: Hoss Man
> Assignee: Hoss Man
> Priority: Major
> Attachments: SOLR-12516.patch, SOLR-12516.patch, SOLR-12516.patch,
> SOLR-12516.patch
>
>
> while simple {{type:range}} facets don't benefit from refinement, because
> every shard returns the same set of buckets, some bugs currently exist when a
> range facet contains sub facets that use refinement:
> # the optional {{other}} buckets (before/after/between) are not considered
> during refinement
> # when using the {{include}} option: if {{edge}} is specified, then the
> refinement of all range buckets mistakenly includes the lower bound of the
> range, regardless of whether {{lower}} was specified.
> ----
> #1 occurs because {{FacetRangeMerger extends
> FacetRequestSortedMerger<FacetRange>}} ... however {{FacetRangeMerger}} does
> not override {{getRefinement(...)}} which means only
> {{FacetRequestSortedMerger.buckets}} is evaluated and considered for
> refinement. The additional, special purpose, {{FacetBucket}} instances
> tracked in {{FacetRangeMerger}} are never considered for refinement.
> #2 exists because of a mistaken in the implementation of {{refineBucket}} and
> how it computes the {{start}} value.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
