tvalentyn commented on issue #36201: URL: https://github.com/apache/beam/issues/36201#issuecomment-3916661993
Dataflow has it's own implementation of pubsub sources/sinks that direct runners, it has been this way for many years and that's not changing. Streaming source wouldn't work in batch mode. > streaming Dataflow behavior might differ from DirectRunner Yeah, ideally behaviors would match. In this case my worry is that the additional functionality in Direct runner would have limited practical application. I think if a Python user must use Ordered Key Publishing, my first suggestion would be to try using Java Pubsub IO through XLang framework. You could try to give a try, check whether it works, and document how to use it as a part of this issue, instead of implementing in Python's direct runner. -- 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]
