[
https://issues.apache.org/jira/browse/HIVE-8613?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14255522#comment-14255522
]
Navis commented on HIVE-8613:
-----------------------------
[~szehon] In test mode, CBO is enabled changing constant(second argument) to
column type and throws exception creating evaluator for the function. It seemed
to be an another issue but before digging deeper, I though checking group-by is
missed by mistake. Before making questions to CBO guys, I wanted to confirm
tests passed.
> percentile_approx raise a comparator error
> ------------------------------------------
>
> Key: HIVE-8613
> URL: https://issues.apache.org/jira/browse/HIVE-8613
> Project: Hive
> Issue Type: Bug
> Affects Versions: 0.12.0
> Reporter: Nicolas Lalevée
> Assignee: Navis
> Attachments: HIVE-8613.1.patch.txt, HIVE-8613.2.patch.txt
>
>
> I have implemented an UADF which wrap around percentile_approx, and I have
> been hit by the following stack trace:
> {noformat}
> Caused by: java.lang.IllegalArgumentException: Comparison method violates its
> general contract!
> at java.util.ComparableTimSort.mergeLo(ComparableTimSort.java:714)
> at java.util.ComparableTimSort.mergeAt(ComparableTimSort.java:451)
> at java.util.ComparableTimSort.mergeCollapse(ComparableTimSort.java:376)
> at java.util.ComparableTimSort.sort(ComparableTimSort.java:182)
> at java.util.ComparableTimSort.sort(ComparableTimSort.java:146)
> at java.util.Arrays.sort(Arrays.java:472)
> at java.util.Collections.sort(Collections.java:155)
> at
> org.apache.hadoop.hive.ql.udf.generic.NumericHistogram.merge(NumericHistogram.java:161)
> {noformat}
> In my use case, with my special UADF, there were Double.NaN
> Implementing a proper double comparaison with Double.compare in
> com.goojet.mapred.hive.udf.NumericHistogram.Coord.compareTo(Object) fixes it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
