[
https://issues.apache.org/jira/browse/KYLIN-6026?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17928303#comment-17928303
]
Guoliang Sun commented on KYLIN-6026:
-------------------------------------
h3. Root Cause
When standalone-deployed Kylin5 performs table sampling, it uses
`sparkSession.ss.catalog.tableExists` to verify if the current table exists.
Since `enableHiveSupport` is enabled, it follows the Hive logic. In a
standalone deployment, the environment lacks Hadoop and Hive, causing Hive
initialization to fail and throw an exception, which leads to sampling failure.
h3. Dev Design
Add a condition: If the table's source type is JDBC (i.e., using a JDBC data
source), skip `calculateViewMetas`.
> JDBC data source table sampling error
> -------------------------------------
>
> Key: KYLIN-6026
> URL: https://issues.apache.org/jira/browse/KYLIN-6026
> Project: Kylin
> Issue Type: Bug
> Affects Versions: 5.0.0
> Reporter: Guoliang Sun
> Priority: Major
> Attachments: image-2025-02-19-15-22-26-359.png
>
>
> !image-2025-02-19-15-22-26-359.png|width=573,height=288!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
