danny0405 commented on a change in pull request #10123:
[FLINK-14665][table-planner-blink] Support computed column for create…
URL: https://github.com/apache/flink/pull/10123#discussion_r344436625
##########
File path:
flink-table/flink-table-planner-blink/src/test/resources/org/apache/flink/table/planner/plan/stream/sql/DagOptimizationTest.xml
##########
@@ -31,13 +31,11 @@
LogicalSink(name=[`default_catalog`.`default_database`.`retractSink`], fields=[b
+- LogicalProject(b=[$1], a=[$0])
+- LogicalUnion(all=[true])
:- LogicalProject(a=[$0], b=[$1], c=[$2])
- : +- LogicalProject(a=[$0], b=[$1], c=[$2])
- : +- LogicalFilter(condition=[LIKE($2, _UTF-16LE'%hello%')])
- : +- LogicalTableScan(table=[[default_catalog,
default_database, MyTable, source: [TestTableSource(a, b, c)]]])
+ : +- LogicalFilter(condition=[LIKE($2, _UTF-16LE'%hello%')])
Review comment:
When we use the `toRel` way to translate the view table, see
`FlinkRelOptTable#toRel` TranslatableTable part, the view can not be expanded
because of the `QueryOperationCatalogViewTable` use a redundant projection to
fix the nullability attribute. This is definitely wrong, we should fix the
nullability from the row type inference. So i remove the hacky projection and
we reduce a projection in the middle plan.
----------------------------------------------------------------
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]
With regards,
Apache Git Services
