[
https://issues.apache.org/jira/browse/HBASE-665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12603096#action_12603096
]
Clint Morgan commented on HBASE-665:
------------------------------------
Yeah, I looked a little closer, and it does not seem that it is a filter issue.
The WhileMatch->StopRow filter should cause each individual store scanner to
stop early.
Still we should be checking filterAllRemaining in the higer-level region filter
(as in my patch). This is where the filterRow is called, so if filters make a
filterAllRemaining decision based on fully assembled rows, then we need the
check there to respect it...
Perhaps its best to just put all the stopRow smarts on the client-side scanner?
> server side scanner doesn't honor stop row
> ------------------------------------------
>
> Key: HBASE-665
> URL: https://issues.apache.org/jira/browse/HBASE-665
> Project: Hadoop HBase
> Issue Type: Bug
> Reporter: Michael Bieniosek
> Fix For: 0.1.3, 0.2.0
>
> Attachments: hbase-665.patch
>
>
> I have a large table. If I create a scanner with a stop row near the
> beginning of the table, the last hasNext call hangs for a while. If I do the
> same with the stop row near the end of the table, the last hasNext call is
> pretty quick.
> I suspect that the server side scanner isn't terminating early, and is
> actually scanning through the whole table returning nothing.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
