[
https://issues.apache.org/jira/browse/LUCENE-7572?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15696309#comment-15696309
]
David Smiley commented on LUCENE-7572:
--------------------------------------
RE LongHashSet: makes sense.
RE PrefixCodedTerms: wow I didn't know it could be more compact than an FST! I
thought the FST reigned supreme here. Would you mind adding some JavaDocs to
PrefixCodedTerms to say this? I do see it definitely has a subset of FST
features.
I was looking at PrefixCodedTerms a bit and I see some minor things that could
be improved. The presence of "delGen" seems out of place having nothing
directly to do with this utility class's job; probably bolted on for
BufferedUpdatesStream's convenience. And RAMFile shouldn't depend on
RAMDirectory for the sole reason of buffer size increase notification; instead
it could take a callback lambda. Of course these are both separate issues from
this one.
> Cache the hashcode of the doc values terms queries
> --------------------------------------------------
>
> Key: LUCENE-7572
> URL: https://issues.apache.org/jira/browse/LUCENE-7572
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Adrien Grand
> Priority: Minor
> Attachments: LUCENE-7572.patch
>
>
> DocValuesNumbersQuery and DocValuesTermsQuery can potentially wrap a large
> number of terms so it would help if we cached their hashcode.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
