[
https://issues.apache.org/jira/browse/HIVE-29627?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18089852#comment-18089852
]
Ruben Q L commented on HIVE-29627:
----------------------------------
Hi [~cyanz]! Sure, I had not started to work on this. Please go ahead :)
> Improve IS_NULL selectivity
> ---------------------------
>
> Key: HIVE-29627
> URL: https://issues.apache.org/jira/browse/HIVE-29627
> Project: Hive
> Issue Type: Improvement
> Reporter: Ruben Q L
> Priority: Major
>
> Currently IS_NULL selectivity computation in FilterSelectivityEstimator does
> not have any special treatment, and just goes over the default computation.
> However, if statistics are available, we could get a much better estimation
> (as it already happens with IS_NOT_NULL) by using numberOfNulls / numRows
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
