[
https://issues.apache.org/jira/browse/CASSANDRA-10487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14950093#comment-14950093
]
Benjamin Lerer commented on CASSANDRA-10487:
--------------------------------------------
The patch for 2.2 is
[here|https://github.com/apache/cassandra/compare/trunk...blerer:10487-2.2] and
the one for 3.0 is
[here|https://github.com/apache/cassandra/compare/trunk...blerer:10487-3.0]
I did not find the change that introduced the regression but it seems that the
problem was caused by the commands using the {{LIMIT}} value to limit the
number of rows retrieved.
The patch for {{2.2}}:
# makes sure that for aggregate queries we do not pass the {{LIMIT}} value to
the commands
# adds extra unit tests
# fix the {{NEWS.txt}}
The patch for {{3.0}} fix the {{NEWS.txt}} and add some extra tests.\\
\\
* The unit test results for 2.2 are
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-10487-2.2-testall/1/]
* The dtest results for 2.2 are
[here|http://cassci.datastax.com/view/Dev/view/blerer/job/blerer-10487-2.2-dtest/1/]
> Regression in select count with limit
> -------------------------------------
>
> Key: CASSANDRA-10487
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10487
> Project: Cassandra
> Issue Type: Bug
> Reporter: Benjamin Lerer
> Assignee: Benjamin Lerer
> Fix For: 2.2.x
>
>
> After CASSANDRA-4914 the {{LIMIT}} clause should limit only the number of
> rows returned to the user not the number of rows used by the aggregates.
> That functionality is currently broken in 2.2.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
