You are viewing a plain text version of this content. The canonical link for it is here.

Posted to solr-dev@lucene.apache.org by "Yonik Seeley (JIRA)" <ji...@apache.org> on 2009/05/14 23:52:45 UTC

[jira] Commented: (SOLR-1169) SortedIntDocSet

    [ https://issues.apache.org/jira/browse/SOLR-1169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12709605#action_12709605 ] 

Yonik Seeley commented on SOLR-1169:
------------------------------------

We need skipping in our DocSets in order to be able to pass them as filters and improve search performance.
One could Sort a HashDocSet every time it's used.... but that's not desirable.

The way Lucene now uses filters, random access performance is no longer important, but being able to efficiently skip is.
Intersection performance is very important to Solr of course, so if we can get SortedIntDocSet performance fast enough then it would make sense for it to replace HashDocSet.

I've already started working on this, and the results look promising.  I'll post a draft soon.

> SortedIntDocSet
> ---------------
>
>                 Key: SOLR-1169
>                 URL: https://issues.apache.org/jira/browse/SOLR-1169
>             Project: Solr
>          Issue Type: Improvement
>            Reporter: Yonik Seeley
>            Assignee: Yonik Seeley
>             Fix For: 1.4
>
>
> A DocSet type that can skip to support SOLR-1165

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.