Sorry, I couldn't find any clear and detailed user guidance other than FLIP in
the official documentation too.
> 2024年10月15日 01:39，Anil Dasari 写道：
>
> Hi Yanquan,
> I've finished reading the Sink FLIPs and am now reviewing some of the sink
> implementations, like TestSinkV2, to better underst
Hi Yanquan,
I've finished reading the Sink FLIPs and am now reviewing some of the sink
implementations, like TestSinkV2, to better understand the flow. I'll write
a new one to experiment with.
Are there flink sink docs/flow daigrams like detailed source implementation
docs like
https://nightlies.ap
Hi, Anil.
For your scenario, I think looking at FLIP-143 first and then FLIP-191
should provide a better understanding. Then you can look at other FLIPs or
specific implementations.
Anil Dasari 于2024年10月15日周二 00:55写道：
> Got it. thanks.
> Sink improvements have many FLIP confluence pages i.e FLIP
Got it. thanks.
Sink improvements have many FLIP confluence pages i.e FLIP-143, 171, 177
and 191. So, Is there a sequence of steps flow charts for better
understanding of the sink process with sink, writer and committer ?
Thanks
On Mon, Oct 14, 2024 at 9:48 AM Yanquan Lv wrote:
> Yeah, TwoPhase
Yeah, TwoPhaseCommittingSink will be removed in Flink 2.0, and it will be
replaced by SupportsCommitter[1] interface, which was introduced in Flink
1.19.
But you can still use TwoPhaseCommittingSink under Flink 2.0, it depends on
your target Flink version, The interfaces of these two APIs are almos
Hi Yanquan,
Thanks for sharing the information.
It appears that TwoPhaseCommittingSink is not available in the flink repo
main branch. it is replaced with Sink, Committer and SinkWritter ?
Thanks
On Mon, Oct 14, 2024 at 1:45 AM Yanquan Lv wrote:
> Hi, Anil.
>
> Iceberg Sink is merged recently i
Hi, Anil.
Iceberg Sink is merged recently in
https://github.com/apache/iceberg/pull/10179#pullrequestreview-2350414880.
From your description, I guess that what you need is a
TwoPhaseCommittingSink[1], the steps you listed can be executed with the
following steps:
> 1. Group data by category
