[
https://issues.apache.org/jira/browse/OAK-5937?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16290487#comment-16290487
]
Thomas Mueller commented on OAK-5937:
-------------------------------------
>> Yes, but the loops for b and c are very short. Only a loops over the whole
>> repository.
> That is the problamatic part and something we know can happen
Sorry, I still don't understand. What exactly is the problematic part, is it
that "a loops over the whole repository"? Yes, it can happen, and as Marcel and
me already wrote, we have the traversal limit that stops such queries. I agree
people should not write expensive queries in general. But I don't see what
would be special about the path restriction.
Maybe you think such queries are run by mistake? The query you gave looks like
a mistake to me: //content/foo/bar. Maybe the user meant
/jcr:root/content/foo/bar. Is this the problem maybe, you want to reject such
queries because they are likely mistakes? Yes this we could do, for example
reject the query //content/foo/bar as a syntax error, and require for example
to use the syntax /jcr:root/*//content/foo/bar in this case. Not sure about the
details, but yes that might make sense.
> Disable query where path restriction is not absolute
> ----------------------------------------------------
>
> Key: OAK-5937
> URL: https://issues.apache.org/jira/browse/OAK-5937
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: query
> Reporter: Chetan Mehrotra
> Priority: Minor
> Fix For: 1.8
>
>
> Query like below cannot be executed in a performant way. We should provide an
> option to reject such queries
> //content/foo/bar
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
