Github user walterddr commented on the issue:
https://github.com/apache/flink/pull/5555
@hequn8128 @fhueske Thanks for the feedback. I have updated the diff to
directly use DistinctAccumulator for filtering and modified the
`generateAggregation` API. Please kindly take another look when you have time.
I have resolved the issue of multiple layer dataview codegen.
In terms of reusing same `DataView` for multiple distinct aggregations
against the same field, I tried to incorporate but there are many assumptions
with single mapping between `AggregateFunction`s and `Accumulator`s that's hard
to deal with. I am planning to continue and improve on it in a separated JIRA,
what do you think?
---
