[
https://issues.apache.org/jira/browse/KYLIN-2483?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15906107#comment-15906107
]
Dayue Gao commented on KYLIN-2483:
----------------------------------
Hi [~mahongbin], I think the overhead could come from record clone in
SortedIteratorMergerWithLimit. I'll try to refine it in KYLIN-2501.
> SortedIteratorMergerWithLimit could be slower when number of total merge rows
> is small
> --------------------------------------------------------------------------------------
>
> Key: KYLIN-2483
> URL: https://issues.apache.org/jira/browse/KYLIN-2483
> Project: Kylin
> Issue Type: Improvement
> Reporter: hongbin ma
> Assignee: hongbin ma
>
> if the pushed down limit is small enough (say less than 100),
> SortedIteratorMergerWithLimit will bring RELATIVELY significant costs. I'm
> adding a new configuration entry called
> kylin.query.merge-sort-partition-results.min-limit (default 100) to fix this
> issue
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
