Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3919#discussion_r117183489
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamGroupWindowAggregate.scala
---
@@ -131,6 +135,19 @@ class DataStreamGroupWindowAggregate(
"non-windowed GroupBy aggregation.")
}
+ val isCountWindow = window match {
+ case TumblingGroupWindow(_, _, size) if isRowCountLiteral(size) =>
true
+ case SlidingGroupWindow(_, _, size, _) if isRowCountLiteral(size) =>
true
+ case _ => false
+ }
+
+ if (isCountWindow && grouping.length > 0 &&
queryConfig.getMinIdleStateRetentionTime < 0) {
+ LOG.warn(
--- End diff --
I think this can be argued in both ways. If we log an error message could
also throw an exception and prevent the program to be executed. However, there
are also other strategies to deal with state cleanup, e.g., by configuring
RocksDB.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
