[
https://issues.apache.org/jira/browse/LUCENE-6198?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14289712#comment-14289712
]
Robert Muir commented on LUCENE-6198:
-------------------------------------
{quote}
It seems this issue is a solution to LUCENE-6032 right?
{quote}
It should be. the motivation here is different, its to try to provide general
performance improvements. If we decide to do it, then I think we should add
support to FilteredQuery/SloppyPhraseQuery/DisjunctionQuery at a minimum, so
that proximity is faster overall regardless of where its placed in the
query/filter tree. Other things are possible, such as implementing for spans so
that they delay reading of positions until there is a match (this would speed
up SpanNearQuery just by itself). I am unsure if we can use it to speed
MinShouldMatch when there is a filter, maybe thats another interesting one.
But yes, I think the idea is that "slow filters" just need to support the
approximation api to work most efficiently regardless of where they are placed.
Even if they return MatchAll as an approximation its better than today.
Obviously, things like booleanfilter or whatever need to implement this api for
that to work.
> two phase intersection
> ----------------------
>
> Key: LUCENE-6198
> URL: https://issues.apache.org/jira/browse/LUCENE-6198
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Robert Muir
> Attachments: LUCENE-6198.patch
>
>
> Currently some scorers have to do a lot of per-document work to determine if
> a document is a match. The simplest example is a phrase scorer, but there are
> others (spans, sloppy phrase, geospatial, etc).
> Imagine a conjunction with two MUST clauses, one that is a term that matches
> all odd documents, another that is a phrase matching all even documents.
> Today this conjunction will be very expensive, because the zig-zag
> intersection is reading a ton of useless positions.
> The same problem happens with filteredQuery and anything else that acts like
> a conjunction.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
