You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Timo Walther (Jira)" <ji...@apache.org> on 2021/11/01 11:44:00 UTC

[jira] [Created] (FLINK-24717) Push partitions before filters

Timo Walther created FLINK-24717:
------------------------------------

             Summary: Push partitions before filters
                 Key: FLINK-24717
                 URL: https://issues.apache.org/jira/browse/FLINK-24717
             Project: Flink
          Issue Type: Sub-task
          Components: Table SQL / Planner
            Reporter: Timo Walther


Currently, we push filters before partitions. This means that a {{applyFilters}} needs to have partition logic to extract the partition predicate. Furthermore, if a {{applyFilters}} consumes all filters (no remaining predicates), the {{applyPartitions}} is never called.

We should execute the {{PushPartitionIntoTableSourceScanRule}} first and check for side effects of this change.

See {{org.apache.flink.table.planner.plan.rules.logical.PushProjectIntoTableSourceScanRuleTest#testMetadataProjectionWithoutProjectionPushDownWhenNotSupportedAndNoneSelected}} for an example of using the new test infrastructure.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)