Hm thanks for pointing this out Brian. It looks like the Java WindowedWordCount example does process an input file while the Java one processes a PubSub topic. Changing the command would be a good quick fix, but I think the best fix would actually be to make the python example mirror the Java one. I'm curious what other dev@ readers think about this though.
(I also wanted to point out Kyle recently discovered BEAM-11944 which was introduced in the website revamp - that's why the code block switchers aren't working on that page) Brian On Sun, Mar 14, 2021 at 2:27 PM Mo Brian <brian...@hotmail.com> wrote: > Hi team, > > > > I’m studying the apache beam from > https://beam.apache.org/get-started/wordcount-example/#windowedwordcount-example > > > > A bit lost on the windowed_wordcount.py and it’s start command: > > > > *windowed_wordcount.py input a pubsub message: * > > lines = p | beam.io.ReadFromPubSub(known_args.input_topic) > > > > *start command provide a file input:* > > python -m apache_beam.examples.windowed_wordcount --input YOUR_INPUT_FILE > \ > > --output_table > PROJECT:DATASET.TABLE \ > > --runner DataflowRunner \ > > --project YOUR_GCP_PROJECT \ > > --temp_location > gs://YOUR_GCS_BUCKET/tmp/ > > > > Should I change the command here? > > > > Thanks > > > > Brian > > > > Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986> for > Windows 10 > > >