Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3386#discussion_r108352300
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamOverAggregate.scala
---
@@ -112,7 +113,12 @@ class DataStreamOverAggregate(
"condition.")
}
case _: RowTimeType =>
- throw new TableException("OVER Window of the EventTime type is not
currently supported.")
+ if (overWindow.lowerBound.isUnbounded &&
overWindow.upperBound.isCurrentRow) {
+ createUnboundedAndCurrentRowEventTimeOverWindow(inputDS)
--- End diff --
Oh, I just recalled why I hadn't added an event-time OVER ROWS windows.
They cannot deal with late data because all results would need to be shifted if
a late record would be inserted.
I would propose to keep this implementation anyway, and just throw an
exception if a user enables handling of late data. What do you think
@hongyuhong and @sunjincheng121?
---
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 feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
