[
https://issues.apache.org/jira/browse/LUCENE-8311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16877849#comment-16877849
]
Adrien Grand commented on LUCENE-8311:
--------------------------------------
It turns out that part of the reason why the patch is making things slower is
that it is moving phrase queries from BlockPostingsEnum, which is specialized
to read freqs and positions only, to BlockImpactsEverythingEnum, which can read
any of docs+freqs, docs+freqs+positios or docs+freqs+positions+offsets. Maybe
we should remove BlockPostingsEnum and have a specialized impacts enum for
positions instead.
The merged impacts look like they have some room for improvement as well. I'm
looking into those issues so that we can then do better testing of LUCENE-8806.
> Leverage impacts for phrase queries
> -----------------------------------
>
> Key: LUCENE-8311
> URL: https://issues.apache.org/jira/browse/LUCENE-8311
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Adrien Grand
> Priority: Minor
> Attachments: LUCENE-8311.patch
>
>
> Now that we expose raw impacts, we could leverage them for phrase queries.
> For instance for exact phrases, we could take the minimum term frequency for
> each unique norm value in order to get upper bounds of the score for the
> phrase.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
