Guowei, thanks a lot. Looks like chainingStrategy default is HEAD. Let me
try.
writer.setChainingStrategy(ChainingStrategy.ALWAYS);
On Sat, Nov 23, 2019 at 7:17 PM Guowei Ma <guowei....@gmail.com> wrote:
> Hi, Steven
>
> 1. The `icebergsink-writer` operator does not chain with the first
> operator is because the "icebergsink-writer" operator 's ChainingStrategy
> is NULL or HEAD. You could verify it by printing writer.getChainingStrategy.
> 2. The two operators use the FORWARD partition if the parallelism of two
> operators is the same and the partitioner is not specified by the user.
>
> Best,
> Guowei
>
>
> Steven Wu <stevenz...@gmail.com> 于2019年11月23日周六 上午5:17写道：
>
>>
>> I have this DAG screenshot from Flink UI.
>> [image: image.png]
>> I am wondering why is the middle "icebergsink-writer" operator not
>> chained with the first operator chain?
>> Or an equivalent question is why is forward partitioner used here?
>>
>> The first operator chain are all map functions after source. The last two
>> operators are added like this.
>> --------------------------
>> dataStream
>> .transform("icebergsink-writer", TypeInformation.of(SomeClass.class),
>> writer)
>> .addSink(committer)
>> .name("icebergsink-committer")
>> .uid("icebergsink-committer")
>> .setParallelism(1);
>>
>> Thanks,
>> Steven
>>
>
