[
https://issues.apache.org/jira/browse/LUCENE-6371?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14564691#comment-14564691
]
Robert Muir commented on LUCENE-6371:
-------------------------------------
By the way, those two go together. Its hard to fix that slow query to do the
right thing and pass along contexts from rewrite while having the leniency, too
scary and undertested.
At the same time, we don't want this scary SpanMultiTermQuery to corrupt all of
our span apis. The optimization in question won't fix fundamental performance
problems with it, and its basically the only one with a problem.
These are the reasons why the original termcontext stuff for spans was a
half-ass implementation, all those problems and very ugly as well, but
relatively simple and contained. The half-ass implementation worked completely
(seeks reduced from 2 to 1) for all normal span queries, just not the crazy
SpanMultiTermQuery. So it was a tradeoff for simplicity, and works well for all
the regular spans use cases, so they were more inline with e.g.
sloppyphrasequery and so on.
> Improve Spans payload collection
> --------------------------------
>
> Key: LUCENE-6371
> URL: https://issues.apache.org/jira/browse/LUCENE-6371
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Paul Elschot
> Assignee: Alan Woodward
> Priority: Minor
> Fix For: Trunk, 5.3
>
> Attachments: LUCENE-6371.patch, LUCENE-6371.patch, LUCENE-6371.patch,
> LUCENE-6371.patch
>
>
> Spin off from LUCENE-6308, see the comments there from around 23 March 2015.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
