lokeshj1703 opened a new issue, #19745:
URL: https://github.com/apache/hudi/issues/19745

   ### Problem
   
   When HoodieStreamer performs sample writes to estimate record size, 
`SparkSampleWritesUtils.doSampleWrites` initializes the sample-writes shadow 
table without specifying a table version. The shadow table therefore defaults 
to `HoodieTableVersion.current()`, even when the user's write config carries a 
different `hoodie.write.table.version` (for example `6` for a table created 
before the Hudi 1.x upgrade).
   
   The `SparkRDDWriteClient` used for the sample write inherits the user's 
write config (and its table version), so it operates on a shadow table whose 
on-disk layout does not match. This version mismatch between the shadow table 
and the write client can surface as version-conditional failures during the 
sample write, causing record-size estimation to silently fail.
   
   ### Expected
   
   The sample-writes shadow table should be created at the same table version 
the write config is configured with, so the shadow table layout matches the 
client operating on it.
   
   ### Environment
   
   - Hudi version: master (1.x)
   - Affects: HoodieStreamer sample-writes record-size estimation on tables 
written at a non-current table version.
   


-- 
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