chamikaramj commented on issue #36496:
URL: https://github.com/apache/beam/issues/36496#issuecomment-3407349100

   Thanks for the investigation @baeminbo.
   
   Sorting was originally added in https://github.com/apache/beam/issues/24361.
   
   I think the issue was that even at Row/schema construction, different calls 
to `TypedSchemaTransformProvider.configurationSchema` may produce different 
schema's (different field orders). So the encoded Row and the schema could end 
up being inconsistent if we do not sort there (or do not add field numbers). 
   
   Java utils you mention do their own sorting, so pre-sorting the schema might 
not be needed there as you mentioned but we might need to use the row at other 
places as well, so we need to maintain a consistent schema.
   
   We can reduce the priority if we do not see an immediate issue here.
   
   


-- 
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