[ 
https://issues.apache.org/jira/browse/SPARK-60901?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Yicong Huang resolved SPARK-60901.
----------------------------------
    Fix Version/s: 4.4.0
       Resolution: Fixed

Issue resolved by pull request 60115
[https://github.com/apache/spark/pull/60115]

> Stop polling in the pipelined UDF input reader
> ----------------------------------------------
>
>                 Key: SPARK-60901
>                 URL: https://issues.apache.org/jira/browse/SPARK-60901
>             Project: Spark
>          Issue Type: Improvement
>          Components: PySpark
>    Affects Versions: 4.4.0
>            Reporter: Yicong Huang
>            Assignee: Yicong Huang
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 4.4.0
>
>
> The pipelined mode added in SPARK-56642 duplicates the worker's processing 
> path in invoke_udf. Keep a single path and move the background input reader 
> into a context manager in worker_util, so pipelined mode only decides whether 
> the input is read ahead in a thread. Also expose 
> ArrowStreamSerializer.flush_per_batch instead of setting a private attribute 
> from worker.py. No behavior change.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to