You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by John Roesler <jo...@confluent.io> on 2019/10/17 19:06:46 UTC
Re: [VOTE] - KIP-213 (new vote) - Simplified and revised.
Hello all,
While writing some new test cases for foreign key joins (as accepted
in KIP-213), I realized that there was an oversight in the review
process: we only proposed to add join methods that take a Materialized
parameter.
This poses an unnecessary burden on users who don't _need_ to
materialize the join result, which could be quite a lot of state and
changelog topic data.
I filed https://issues.apache.org/jira/browse/KAFKA-9058 and then
https://github.com/apache/kafka/pull/7541 to fix it. I've updated the
KIP page accordingly as well.
This change consists solely of adding the missing method overloads,
and doesn't change the behavior or semantics of the operation at all.
Thanks,
-John
Re: [VOTE] - KIP-213 (new vote) - Simplified and revised.
Posted by Guozhang Wang <wa...@gmail.com>.
Thanks for the catch John, the JIRA description looks good!
On Sat, Oct 19, 2019 at 11:43 PM Matthias J. Sax <ma...@confluent.io>
wrote:
> Thanks John!
>
> Good catch!
>
> On 10/17/19 12:11 PM, Adam Bellemare wrote:
> > Awesome. Thanks John for fixing this!
> >
> > On Thu, Oct 17, 2019 at 3:07 PM John Roesler <jo...@confluent.io> wrote:
> >
> >> Hello all,
> >>
> >> While writing some new test cases for foreign key joins (as accepted
> >> in KIP-213), I realized that there was an oversight in the review
> >> process: we only proposed to add join methods that take a Materialized
> >> parameter.
> >>
> >> This poses an unnecessary burden on users who don't _need_ to
> >> materialize the join result, which could be quite a lot of state and
> >> changelog topic data.
> >>
> >> I filed https://issues.apache.org/jira/browse/KAFKA-9058 and then
> >> https://github.com/apache/kafka/pull/7541 to fix it. I've updated the
> >> KIP page accordingly as well.
> >>
> >> This change consists solely of adding the missing method overloads,
> >> and doesn't change the behavior or semantics of the operation at all.
> >>
> >> Thanks,
> >> -John
> >>
> >
>
>
--
-- Guozhang
Re: [VOTE] - KIP-213 (new vote) - Simplified and revised.
Posted by "Matthias J. Sax" <ma...@confluent.io>.
Thanks John!
Good catch!
On 10/17/19 12:11 PM, Adam Bellemare wrote:
> Awesome. Thanks John for fixing this!
>
> On Thu, Oct 17, 2019 at 3:07 PM John Roesler <jo...@confluent.io> wrote:
>
>> Hello all,
>>
>> While writing some new test cases for foreign key joins (as accepted
>> in KIP-213), I realized that there was an oversight in the review
>> process: we only proposed to add join methods that take a Materialized
>> parameter.
>>
>> This poses an unnecessary burden on users who don't _need_ to
>> materialize the join result, which could be quite a lot of state and
>> changelog topic data.
>>
>> I filed https://issues.apache.org/jira/browse/KAFKA-9058 and then
>> https://github.com/apache/kafka/pull/7541 to fix it. I've updated the
>> KIP page accordingly as well.
>>
>> This change consists solely of adding the missing method overloads,
>> and doesn't change the behavior or semantics of the operation at all.
>>
>> Thanks,
>> -John
>>
>
Re: [VOTE] - KIP-213 (new vote) - Simplified and revised.
Posted by Adam Bellemare <ad...@gmail.com>.
Awesome. Thanks John for fixing this!
On Thu, Oct 17, 2019 at 3:07 PM John Roesler <jo...@confluent.io> wrote:
> Hello all,
>
> While writing some new test cases for foreign key joins (as accepted
> in KIP-213), I realized that there was an oversight in the review
> process: we only proposed to add join methods that take a Materialized
> parameter.
>
> This poses an unnecessary burden on users who don't _need_ to
> materialize the join result, which could be quite a lot of state and
> changelog topic data.
>
> I filed https://issues.apache.org/jira/browse/KAFKA-9058 and then
> https://github.com/apache/kafka/pull/7541 to fix it. I've updated the
> KIP page accordingly as well.
>
> This change consists solely of adding the missing method overloads,
> and doesn't change the behavior or semantics of the operation at all.
>
> Thanks,
> -John
>