[
https://issues.apache.org/jira/browse/LUCENE-9584?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17220608#comment-17220608
]
hackerwin7 edited comment on LUCENE-9584 at 11/18/20, 10:57 AM:
----------------------------------------------------------------
I upload a patch based on 7.7.0, just check the result iterator before new
ConstantScoreScorer; but I'm not sure should we use return null to terminate
more early in BooleanWeight.scorerSupplier() just let the subScorer == null.
I believe that in PointInSetQuery.scorer()
reader.getPointValues(field) == null *is equal to* reader intersected result
is empty
this two case should both return null to terminate more early.
was (Author: hackerwin7):
I upload a patch based for 7.7.0, just check the result iterator before new
ConstantScoreScorer; but I'm not sure should we use return null to terminate
more early in BooleanWeight.scorerSupplier() just let the subScorer == null.
I believe that in PointInSetQuery.scorer()
reader.getPointValues(field) == null *is equal to* reader intersected result
is empty
this two case should both return null to terminate more early.
> PointInSetQuery does not terminate early if result iterator has no docs
> -----------------------------------------------------------------------
>
> Key: LUCENE-9584
> URL: https://issues.apache.org/jira/browse/LUCENE-9584
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/search
> Affects Versions: 7.7.3, 8.6.3
> Reporter: hackerwin7
> Priority: Major
> Labels: performance
> Attachments: LUCENE-7.7.0-PointInSetQuery_terminate_early.patch
>
>
> Today, in a point in set query after BKD intersect we get a DocIdSetBuilder
> result, if result's iterator have no docs, then the PointInSetQuery still
> create ConstantScoreScorer with an empty DocIdSetIterator.
> In a Boolean Query, such as query = subQuery1 AND subQuery2 AND subQuery3
> .... subQueryN
> if subQuery1 is a PointInSetQuery and get an empty result iterator, and
> subsequent subQuery2 ~ subQueryN would still evaluate to call build scorer,
> this is an unnecessary cost for this query if subQuery1 have already got an
> empty result iterator .
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
