ATHARVA262005 commented on issue #37371:
URL: https://github.com/apache/beam/issues/37371#issuecomment-3791398694

   Hi @shunping , I have an update! 🚀
   
   I investigated the AssertionError in Batch mode and found the root cause. It 
appears that DataflowRunner.java was not applying the GroupIntoBatches (and 
ShardedKey) overrides when the Runner V2 / Unified Worker experiments were 
active. This caused the pipeline to fall back to a default stateful ParDo 
implementation that ignored the user-defined batch limits, leading to the 
failed assertions.
   
   I have updated PR #37402 with:
   
   Core Fix: Updated the translation logic in DataflowRunner.java to explicitly 
apply these overrides for the V2 Batch path.
   
   Regression Test: Added 
testBatchGroupIntoBatchesWithShardedKeyOverrideCountV2 to 
DataflowRunnerTest.java, which forces the V2/Unified Worker flags to ensure 
this remains fixed.
   
   All tests, including the new regression test, are now passing locally in 
Batch mode. I'd love your review on this fix!


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to