You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "SimonAries (Jira)" <ji...@apache.org> on 2022/09/15 06:31:00 UTC

[jira] [Commented] (SPARK-40441) With PANDAS_UDF, data from tasks on the same physical node is aggregated into one task execution, resulting in concurrency not being fully utilized

    [ https://issues.apache.org/jira/browse/SPARK-40441?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17605134#comment-17605134 ] 

SimonAries commented on SPARK-40441:
------------------------------------

I hope the big guys give me some Pointers

> With PANDAS_UDF, data from tasks on the same physical node is aggregated into one task execution, resulting in concurrency not being fully utilized
> ---------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-40441
>                 URL: https://issues.apache.org/jira/browse/SPARK-40441
>             Project: Spark
>          Issue Type: Bug
>          Components: Pandas API on Spark, PySpark, Scheduler
>    Affects Versions: 2.4.4
>            Reporter: SimonAries
>            Priority: Major
>         Attachments: image-2022-09-15-14-28-04-332.png, image-2022-09-15-14-29-35-004.png
>
>
> This caused the data skew to be very serious, and I did repartition operation before executing the data
> !image-2022-09-15-14-29-35-004.png!
> !image-2022-09-15-14-28-04-332.png!



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org