scwhittle opened a new issue, #22723:
URL: https://github.com/apache/beam/issues/22723

   ### What happened?
   
   The python sdk BatchElements specializes the global window with 
_GlobalWindowsBatchingDoFn
   
   This builds up batches and outputs them in two possible places:
   - during element processing if the batch is deemed complete, this uses the 
current element's timestamp for the output
   - during finish_bundle if there was an unfinished batch, this uses 
GlobalWindows.windowed_value which uses the minimum timestamp
   
   It seems like this should be made to be consistent by either:
   - always outputtting with maximum timestamp. This matches the behavior of 
the non-global window case where the window end timestamp is used.
   - keep track of the minimum timestamp of elements in the input batch and use 
that for the output timestamp
   
   ### Issue Priority
   
   Priority: 2
   
   ### Issue Component
   
   Component: sdk-py-core


-- 
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