I feel like this is related to https://www.youtube.com/watch?v=7lJyq1hw_KI#t=18m
> On May 17, 2021, at 3:36 PM, Pierre Oberholzer <pierre.oberhol...@gmail.com> > wrote: > > Dear Beam Community, > > We would like to run a streaming job from Pub/Sub to BigQuery that handles > schema updates "smoothly" (i.e. without having to restart a new job). Any > suggestion on a suitable method/architecture to achieve this ? > > We found the below unresolved question [1] on Stack Overflow where the Beam > user list was referred to. > > Thanks a lot for your support ! > > [1] > https://stackoverflow.com/questions/60496227/how-can-i-write-streaming-dataflow-pipelines-that-support-schema-evolution > > <https://stackoverflow.com/questions/60496227/how-can-i-write-streaming-dataflow-pipelines-that-support-schema-evolution> > > -- > Pierre
