Au-Miner commented on code in PR #25789:
URL: https://github.com/apache/flink/pull/25789#discussion_r1883583419
##########
flink-table/flink-table-planner/src/main/java/org/apache/flink/table/planner/plan/nodes/exec/stream/StreamExecGroupTableAggregate.java:
##########
@@ -120,16 +119,16 @@ protected Transformation<RowData> translateToPlanInternal(
final AggsHandlerCodeGenerator generator =
new AggsHandlerCodeGenerator(
- new CodeGeneratorContext(
- config,
planner.getFlinkContext().getClassLoader()),
- planner.createRelBuilder(),
-
JavaScalaConversionUtil.toScala(inputRowType.getChildren()),
- // TODO: heap state backend do not copy key
currently,
- // we have to copy input field
- // TODO: copy is not need when state backend
is rocksdb,
- // improve this in future
- // TODO: but other operators do not copy this
input field.....
- true)
+ new CodeGeneratorContext(
+ config,
planner.getFlinkContext().getClassLoader()),
+ planner.createRelBuilder(),
+
JavaScalaConversionUtil.toScala(inputRowType.getChildren()),
+ // TODO: heap state backend do not copy key currently,
Review Comment:
Currently, sync and asynchronous use different versions of state objects. In
consideration of
https://github.com/apache/flink/commit/bccace0f4b233e10279c7d95e009ae6aadad5ae8,
the design concept is to design another set of corresponding files and reuse
the code logic as much as possible in `Helper`
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
