[
https://issues.apache.org/jira/browse/JCR-1337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12562409#action_12562409
]
Ard Schrijvers commented on JCR-1337:
-------------------------------------
"Because most DocIds are PlainDocIds, which do not need to be resolved, the
pre-warming is quite effective IMO"
Yes I already figured this out. It is indeed effective as long as the indexes
have not very many deleted/updated nodes (when the number of parent nodes in
foreign segment grows). I agree this is an easy first iteration optimization.
Also, when the number of foreign parent might get very high, users good choose
to recreate the index by deleting it (if it is an option ofcourse :-) )
I think that when Lucene would have support for uid in the future, we might
also be able to pre-warm the parents when they are in different indexes.
+1 for your patch (though i have not yet had time to test performance gain.
Still try to find some time for it soon)
> Optimize first execution queries for DescendantSelfAxisWeight/ChildAxisQuery
> ----------------------------------------------------------------------------
>
> Key: JCR-1337
> URL: https://issues.apache.org/jira/browse/JCR-1337
> Project: Jackrabbit
> Issue Type: Improvement
> Components: query
> Affects Versions: 1.4
> Reporter: Ard Schrijvers
> Fix For: 2.0
>
> Attachments: JCR-1337.patch, JCR-1337.patch, JCR-1337.patch
>
>
> The first execution of a query involving
> DescendantSelfAxisWeight/ChildAxisQuery is slow. Consecutive queries are
> faster because the hierarchy is cached
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
