Github user fhueske commented on the issue:
https://github.com/apache/flink/pull/3783
Hi @huawei-flink ,
I'll close this PR later today when merging #.
# adds runtime support for distinct aggregation and follows a similar
approach as this PR. However, it leverages
Github user fhueske commented on the issue:
https://github.com/apache/flink/pull/3783
Not yet, but AFAIK is @twalthr working on that.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this
Github user stefanobortoli commented on the issue:
https://github.com/apache/flink/pull/3783
@fhueske thanks for the comments. Did we include the latest calcite already?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If
Github user huawei-flink commented on the issue:
https://github.com/apache/flink/pull/3783
@fhueske , this is the PR with the code generated distinct aggregation for
OVER. You mentioned that the value of the aggregation should be a Row, but what
is kept in the distinct state is just
Github user rtudoran commented on the issue:
https://github.com/apache/flink/pull/3783
@fhueske thanks for the clarification. I think it is good also to have the
solution for the over windows :)
I also wanted to ask you about the calcite and DISTINCT/DIST syntax. What
do
Github user fhueske commented on the issue:
https://github.com/apache/flink/pull/3783
Hi @huawei-flink, I did not say that we need to move the state out of the
code-gen'd function. We can and should leave the PR as it is.
However, we cannot use this code for distinct group
Github user huawei-flink commented on the issue:
https://github.com/apache/flink/pull/3783
@rtudoran @fhueske the first implementation I made was with the state in
the ProcessFunction without code generated aggregation function. Second, I
pushed a branch with the state in the process
Github user rtudoran commented on the issue:
https://github.com/apache/flink/pull/3783
@fhueske @stefanobortoli
Regarding the options for solving the distinct. From my point of view the
previous approach worked:
- we keep in the processFunctions a state for each field that is
Github user rtudoran commented on the issue:
https://github.com/apache/flink/pull/3783
@fhueske @stefanobortoli
I recently fixed in Calcite the problem of porting distinct flag to the the
over. This was merged in the master. Hence it is a matter of when flink will
get the new
Github user fhueske commented on the issue:
https://github.com/apache/flink/pull/3783
Thanks for this PR @huawei-flink!
I think I made a mistake when I suggested to use the code-gen'd functions
with registered `MapState` to compute distinct window aggregations. Originally,
10 matches
Mail list logo