[
https://issues.apache.org/jira/browse/SPARK-22164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16202081#comment-16202081
]
Zhenhua Wang edited comment on SPARK-22164 at 10/12/17 3:21 PM:
----------------------------------------------------------------
[~ron8hu] I don't think we will use histogram in group-by estimation, ndv can
be updated through propagation based on histogram (if any), we only need ndv
here. Let's close this?
was (Author: zenwzh):
[~ron8hu] I don't think histogram can help with group-by, let's close this?
> support histogram in estimating the cardinality of aggregate (or group-by)
> operator
> -----------------------------------------------------------------------------------
>
> Key: SPARK-22164
> URL: https://issues.apache.org/jira/browse/SPARK-22164
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 2.2.0
> Reporter: Ron Hu
>
> Histogram is effective in dealing with skewed distribution. After we generate
> histogram information for column statistics, we need to adjust aggregate (or
> group-by) cardinality estimation based on equi-height histogram information.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
