Hi Flink Devs:

I am writing a new Iceberg Sink based on the SinkV2 Flink interface [1]. I
have a question regarding the 3 operators that are created behind the
scenes by the framework:

   1. How can I set the UID on these 3 operators? Ideally, what I would
   like, is to propagate the UID of the sink (which will be specified
   optionally by the user) and appending the same UID to each operator (ie:
   $uidPrefix-writer and $uidPrefix-global-comitter and $uidPrefix-commiter).
   I came across the `CustomSinkOperatorUidHashes` class, but this class deals
   directly with the UID hashes, and I feel there should be a way to just
   specify the UID and not the hashes themselves.
   2. How can I set the parallelism of these operators? Can I do it? Or
   would they inherit the parallelism of the sink?


Thanks a lot

-Rodrigo


[1] https://github.com/apache/iceberg/pull/10179
<https://github.com/apache/iceberg/pull/10179#>

Reply via email to