Hi Felix,
when calling dataStream.addSink(new BucketingSink()) it returns a
DataStreamSink object. On this object you can set the parallelism via
setParallelism.
Cheers,
Till
On Mon, Aug 21, 2017 at 7:04 PM, Felix Cheung
wrote:
> Hi,
>
> I'm implementing a custom sink. The job is reading a D
Hi,
I'm implementing a custom sink. The job is reading a DataStream into this
custom sink.
I'd like to be able to maximize the parallelism to use all available slots in
the cluster, but to
write to a smaller sets of files in the final output.
When I implement this sink with DataStream.writeAsTe