Spark runner status is correctly reflected on the Beam website https://beam.apache.org/documentation/runners/capability-matrix/ We don't support python on Spark runner yet. However we have added a lot of components which will be used to support Python on Spark runner as part of building Flink runner. Let me know if you are interested in contributing to expedite spark runner :)
On Mon, Oct 15, 2018 at 3:34 PM Wei Yan <[email protected]> wrote: > Great. Thanks, Ankur. I'll give a try. > BTW, do you happen to know the progress on spark runner support side? > > -Wei > > On Mon, Oct 15, 2018 at 3:29 PM Ankur Goenka <[email protected]> wrote: > >> Hi Wei, >> >> We have made substantial progress towards supporting Python on Flink >> runner though its not production ready yet. >> We will soon be updating the beam website to better reflect Flink support >> but for now here is how you can run python wordcount on Flink >> https://beam.apache.org/contribute/portability/#status >> >> Thanks, >> Ankur >> >> On Mon, Oct 15, 2018 at 3:21 PM Wei Yan <[email protected]> wrote: >> >>> Hi, folks, >>> >>> Would like a quick check on Python SDK support for Flink runner. I saw >>> subtasks of BEAM-2889 <https://jira.apache.org/jira/browse/BEAM-2889> >>> have been resolved. But the getting_started >>> <https://beam.apache.org/get-started/quickstart-py/> page shows python >>> SDK only supports local and dataflow. So would like to check progress here, >>> or any other work left here. >>> >>> -Wei >>> >>
