wuchong commented on a change in pull request #14154:
URL: https://github.com/apache/flink/pull/14154#discussion_r528522574
##########
File path:
flink-table/flink-table-planner-blink/src/test/resources/org/apache/flink/table/planner/plan/stream/sql/TableScanTest.xml
##########
@@ -556,10 +577,10 @@ LogicalProject(a=[$1], b=[$2], c=[$3])
<Resource name="planAfter">
<![CDATA[
Calc(select=[a, b, c], where=[>(a, 1)], changelogMode=[I,UB,UA,D])
-+- WatermarkAssigner(rowtime=[ts], watermark=[-(ts, 1000:INTERVAL SECOND)],
changelogMode=[I,UB,UA,D])
- +- Calc(select=[id, a, +(a, 1) AS b, c, TO_TIMESTAMP(c) AS ts],
changelogMode=[I,UB,UA,D])
- +- ChangelogNormalize(key=[id], changelogMode=[I,UB,UA,D])
- +- Exchange(distribution=[hash[id]], changelogMode=[UA,D])
++- ChangelogNormalize(key=[id], changelogMode=[I,UB,UA,D])
+ +- Exchange(distribution=[hash[id]], changelogMode=[UA,D])
+ +- WatermarkAssigner(rowtime=[ts], watermark=[-(ts, 1000:INTERVAL
SECOND)], changelogMode=[UA,D])
+ +- Calc(select=[id, a, +(a, 1) AS b, c, TO_TIMESTAMP(c) AS ts],
changelogMode=[UA,D])
Review comment:
I think this can be an improvement. Because `ChangelogNormalize` along
with the upsert source is introduced in this version, there is no regressions.
We can pull out the `+(a, 1)` to be after the `ChangelogNormalize` in the
future. I have created https://issues.apache.org/jira/browse/FLINK-20289 to
track this improvement.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
