scwhittle commented on code in PR #22834:
URL: https://github.com/apache/beam/pull/22834#discussion_r953465739
##########
sdks/python/apache_beam/transforms/util.py:
##########
@@ -541,15 +541,15 @@ def process(self, element):
self._running_batch_size += self._element_size_fn(element)
if self._running_batch_size >= self._target_batch_size:
with self._batch_size_estimator.record_time(self._running_batch_size):
- yield self._batch
+ yield window.GlobalWindows.windowed_value_at_end_of_window(self._batch)
Review Comment:
This makes it consistent, but do you think that it will cause issues?
By doing this we're losing insight into the watermark, any produced element
from the batch will automatically inherit the end of global window timestamp.
Should we allow configuring this behavior with a TimestampCombiner?
--
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]
