Xpray commented on a change in pull request #6455: [FLINK-9997] [table] Improve
Expression Reduce
URL: https://github.com/apache/flink/pull/6455#discussion_r222677089
##########
File path:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/api/BatchTableEnvironment.scala
##########
@@ -480,11 +480,20 @@ abstract class BatchTableEnvironment(
normalizedPlan
}
- // 4. optimize the physical Flink plan
+ // 4. optimize the logical Flink plan
+ val flinkLogicalOptRuleSet = getFlinkLogicalOptRuleSet
Review comment:
@twalthr You're right, ReduceExpressionRule.CALC should add to logical opt
phase, but I tried it and it result to java.lang.OutOfMemoryError when I test
`testReduceDeterministicUDF`. I believe the calc rules (`FilterCalcMergeRule,
ProjectCalcMergeRule, FilterToCalcRule, ProjectToCalcRule, CalcMergeRule`)
interact with each other.
If adding a new phase after, there's only calc node, the search space is
quite small than in logical opt phase.
I think adding phase is necessary as the number of rule increasing.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
