[
https://issues.apache.org/jira/browse/SOLR-13892?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jason Gerlowski updated SOLR-13892:
-----------------------------------
Attachment: SOLR-13892.patch
Status: Open (was: Open)
Attaching a patch with a first pass at this:
*Design*
The bulk of this code lives in PostFIlterJoinQuery. On post-filter creation,
we run the {{from}} query using a Collector implementation that populates a
terms bitset ({{fromOrdBitSet}}) as each {{from}} match is found. We then
build a corresponding bitset ({{toOrdBitSet}}) for the {{to}} field by
iterating through the ordinals set in {{fromOrdBitSet}} and checking whether
each set term exists in the {{to}} field. At post-filter execution time, the
postfilter just looks at the individual doc and checks whether it's value in
the {{to}} field matches ones of the bits set in {{toOrdBitSet}}.
In some test scenarios this increases join performance up to 5x over standard
execution. Since bitsets are used, the postfilter scales well as {{from}}
query matches increase. Like most post-filter implementations, it starts to
get slower than standard execution as the {{to}} query results increase (since
postfilters are invoked on each doc matched, this is to be expected).
*Current Limitations*
* only possible on fields with docValues
* only tested with string fields.
* currently only supports single-value fields.
*TODO*
* Tests
* Documentation
* Maybe merge PostFilterJoinQuery in with the existing JoinQuery implementation.
* add support for multivalue fields.
> Add postfilter support to {!join} queries
> -----------------------------------------
>
> Key: SOLR-13892
> URL: https://issues.apache.org/jira/browse/SOLR-13892
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Components: query parsers
> Affects Versions: master (9.0)
> Reporter: Jason Gerlowski
> Priority: Major
> Attachments: SOLR-13892.patch
>
>
> The JoinQParserPlugin would be a lot performant in many use-cases if it could
> operate as a post-filter, especially when doc-values for the involved fields
> are available.
> With this issue, I'd like to propose a post-filter implementation for the
> {{join}} qparser.
--
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
