[
https://issues.apache.org/jira/browse/LUCENE-4224?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13415572#comment-13415572
]
Martijn van Groningen commented on LUCENE-4224:
-----------------------------------------------
The code specialization should give a performance improvement when from terms
are scattered across the terms enum of the to field (in other words when there
gaps between the terms). In the case the matching terms are all next to each
other this code specialization shouldn't buy any additional speed. At least
this is what I'm thinking now.
bq. Note that, both these scorers return docids out of order i think? I don't
think the single-valued one should implement advance(), it should throw UOE?
Hmmm... I think so as well, since it iterates over the terms and per term emits
doc ids. So it can never be in order.
> Simplify MultiValuedCase in TermsIncludingScoreQuery
> ----------------------------------------------------
>
> Key: LUCENE-4224
> URL: https://issues.apache.org/jira/browse/LUCENE-4224
> Project: Lucene - Java
> Issue Type: Task
> Reporter: Robert Muir
> Attachments: LUCENE-4224.patch
>
>
> While looking at LUCENE-4214, i was trying to wrap my head around what this
> is doing...
> I think the code specialization in the multivalued scorer doesn't buy us any
> additional speed? At least according to my benchmarks?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
