kennknowles commented on code in PR #36285:
URL: https://github.com/apache/beam/pull/36285#discussion_r2411270960
##########
sdks/python/apache_beam/transforms/trigger.py:
##########
@@ -933,6 +980,13 @@ def to_runner_api(self, context):
def has_ontime_pane(self):
return any(t.has_ontime_pane() for t in self.triggers)
+ def get_continuation_trigger(self):
+ return Repeatedly(
Review Comment:
Yea, it's not an exact science. There's no real correctness criteria for
continuation trigger except "don't hold up data that is already triggered". And
the only reason we don't make all of them `Repeatedly(Always)` is for the
corner case of aligned processing time where the user might be surprised if a
downstream aggregation had many more outputs because it fired right away
instead of waiting for everything aligned to the same processing time. TBH even
then it is sort of meh.
--
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]
