[
https://issues.apache.org/jira/browse/KYLIN-6010?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17928276#comment-17928276
]
Guoliang Sun commented on KYLIN-6010:
-------------------------------------
h3. Root Cause
In the build code, the following content causes the issue: even though the
global `gluten.enable` is set, the `NativeScan hive xxx.xxx` is not generated
during the snapshot build step, leading to subsequent execution failure.
!image-2025-02-19-14-26-58-795.png|width=731,height=115!
When setting `gluten.enable=false` globally, the same exception still occurs
because `spark.plugins=org.apache.gluten.GlutenPlugin` is configured, causing
the execution to enter the Gluten logic and throw an exception.
> If there is a `count distinct` in the view table definition, an error occurs
> when building the snapshot
> -------------------------------------------------------------------------------------------------------
>
> Key: KYLIN-6010
> URL: https://issues.apache.org/jira/browse/KYLIN-6010
> Project: Kylin
> Issue Type: Bug
> Affects Versions: 5.0.0
> Reporter: Guoliang Sun
> Priority: Major
> Attachments: image-2025-02-19-14-26-58-795.png
>
>
> see following
> {code:java}
> java.lang.UnsupportedOperationException: count distinct does not have
> non-columnar implementation
> at
> org.apache.spark.sql.catalyst.expressions.aggregate.CountDistinct.initialValues$lzycompute(CountDistinct.scala:40)
> ~[gluten.jar]
> at
> org.apache.spark.sql.catalyst.expressions.aggregate.CountDistinct.initialValues(CountDistinct.scala:38)
> ~[gluten.jar]
> at
> org.apache.spark.sql.catalyst.expressions.aggregate.CountDistinct.initialValues(CountDistinct.scala:27)
> ~[gluten.jar] {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
