cshuo commented on PR #18103:
URL: https://github.com/apache/hudi/pull/18103#issuecomment-3871569955

   > @cshuo thx for reviewing, all comments are resolved except `transient` 
comment.
   > 
   > If we remove `transient`, jobs will recover failed because of the state 
incompatibility in case `read.splits.limit` enabled.
   > 
   > Exception as below:
   > 
   > <img alt="截屏2026-02-09 11 49 00" width="2000" height="1022" 
src="https://private-user-images.githubusercontent.com/33287603/546951248-a932bd34-f3d2-4140-beac-272ec412714d.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.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._yTxfHB3e8pJiPm5d18Ypt7Y-g8zRe9WIXP2PzYM_AI";>
   
   Got it.  `inputSplitsState` in StreamReadMonitoringFunction uses the 
fallback kryo serializer for the state serialization, which doesn't schema 
evolution for state. But in StreamReadOperator, there is a `inputSplitsState` 
which is a list state for  `MergeOnReadInputSplit` using JavaSerializer as 
state serializer, and it supports state schema evolution. 
   So it's kind of ambiguous to add a new field to `MergeOnReadInputSplit` with 
`transient`. Maybe we can just extract partition path from the file path 
directly, since it doesn't incur much extra costs. 


-- 
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