[
https://issues.apache.org/jira/browse/SPARK-43131?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dongjoon Hyun updated SPARK-43131:
----------------------------------
Fix Version/s: (was: 3.5.0)
(was: 3.4.1)
(was: 3.3.3)
> Add labels to identify UDF
> ---------------------------
>
> Key: SPARK-43131
> URL: https://issues.apache.org/jira/browse/SPARK-43131
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.2.0
> Reporter: Ahmed Hussein
> Priority: Major
>
> I wish there is a convenient way to report the UDFs executed by spark
> application. This is a useful information to analyze the performance and
> possible optimizations.
> Given an application eventlog, I'm not seeing anything in the plans that can
> easily identify UDFs.
> I see that SPARK-34388 added "source" to the expressionInfo, but this
> extra-info is not propagated to the "explain" or the eventlog.
> I suggest that UDF shows in the eventlog as something like
> {code:scala}
> explain sql("SELECT myUdf1(myUdf2(1))"))
> myUdf1#scalaUDF(myUdf2#scalaUDF(1))
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
