[
https://issues.apache.org/jira/browse/LUCENE-10120?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17437368#comment-17437368
]
Greg Miller commented on LUCENE-10120:
--------------------------------------
Ah [~ChrisLu] I just saw you posted a PR. Looks like we took similar approaches
(see the patch file I attached). I'm not sure we necessarily need to handle the
complexity of docs being added out-of-order though? Unless you've got some
other applications in mind where this might be true. During caching, the docs
will always be visited in-order, so I think some of this can be simplified
down. What do you think? Feel free to incorporate the ideas from my patch file
into your PR if you think they make sense. I'd be curious if benchmarks reveal
any improvements by taking this approach as well. Is that something you're
going to look into?
Thanks again for suggesting this and working on it!
> Lazy initialize FixedBitSet in LRUQueryCache
> --------------------------------------------
>
> Key: LUCENE-10120
> URL: https://issues.apache.org/jira/browse/LUCENE-10120
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/search
> Affects Versions: main (10.0)
> Reporter: Lu Xugang
> Priority: Major
> Attachments: 1.png, LUCENE-10120.patch
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Basing on the implement of collecting docIds in DocsWithFieldSet, may be we
> could do similar way to cache docIdSet in
> *LRUQueryCache#cacheIntoBitSet(BulkScorer scorer, int maxDoc)* when docIdSet
> is density.
> In this way , we do not always init a huge FixedBitSet which sometime is not
> necessary when maxDoc is large
>
>
>
--
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
