[
https://issues.apache.org/jira/browse/LUCENE-10460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17506528#comment-17506528
]
Julie Tibshirani commented on LUCENE-10460:
-------------------------------------------
It indeed seems okay to use a simple DocIdSetIterator#range in this case. I'm
wondering about the motivation for specializing this case though, especially
since the logic is already pretty complex. Have you seen it make a latency
difference when there are missing values? In the case with no missing values I
don't think it will help much, since iterating dense doc values is already
optimized (see DenseNumericDocValues).
> Delegating DocIdSetIterator could be replaced to DocIdSetIterator#range(int
> minDoc, int maxDoc) in IndexSortSortedNumericDocValuesRangeQuery
> --------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: LUCENE-10460
> URL: https://issues.apache.org/jira/browse/LUCENE-10460
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Lu Xugang
> Priority: Trivial
> Time Spent: 10m
> Remaining Estimate: 0h
>
> While taking advantage of of index sort In
> IndexSortSortedNumericDocValuesRangeQuery, if MissingValue disabled, all
> Documents between a range of firstDoc and lastDoc must contain DocValues. So
> In BoundedDocSetIdIterator#advance(int), the delegating DocIdSetIterator
> could be replaced to DocIdSetIterator#range(int minDoc, int maxDoc)?
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
