My understanding is that the minimal_wordcount isn't intended to be run on any distributed runner, or serve as a minimal _production_ pipeline.
It's to serve as a toy example to what a beam pipeline looks like, and introduce the model concepts, which can be learned separately. This may not be adequately documented at present. I agree that it doesn't serve as the best tutorial/introduction as of yet. Someone would need to spend the time to improve this. That said, the other wordcount examples are closer to production style pipelines, so these kinds of updates would be reasonable for them. [ Full content available at: https://github.com/apache/beam/pull/6386 ] This message was relayed via gitbox.apache.org for [email protected]
