[FLINK-12288] [table-planner-blink] Bump Calcite dependency to 1.19.0 in blink planner#8234
[FLINK-12288] [table-planner-blink] Bump Calcite dependency to 1.19.0 in blink planner#8234wuchong wants to merge 2 commits intoapache:masterfrom
Conversation
|
Thanks a lot for your contribution to the Apache Flink project. I'm the @flinkbot. I help the community Review Progress
Please see the Pull Request Review Guide for a full explanation of the review process. DetailsThe Bot is tracking the review progress through labels. Labels are applied according to the order of the review items. For consensus, approval by a Flink committer of PMC member is required Bot commandsThe @flinkbot bot supports the following commands:
|
… in blink planner
| +- LogicalProject(i=[$0], k=[$1]) | ||
| +- LogicalTableScan(table=[[MyTable4, source: [TestTableSource(i, k)]]]) | ||
| +- LogicalProject(i=[$0], k=[$1]) | ||
| +- LogicalTableScan(table=[[MyTable4, source: [TestTableSource(i, k)]]]) |
There was a problem hiding this comment.
It seems that Calcite removes LogicalSort in the initial RelNode tree. And it makes sense.
|
Didn't any plans be changed in flink-planner? I note all changed plans are in blink-planner. |
|
Hi @godfreyhe , I only changed Calcite version in blink planner. Flink planner is still using 1.18.0. |
| <artifactId>calcite-core</artifactId> | ||
| <!-- When updating the Calcite version, make sure to update the dependency exclusions --> | ||
| <version>1.18.0</version> | ||
| <version>1.19.0</version> |
There was a problem hiding this comment.
@wuchong Has Calcite changed any dependencies in the last release? Please verify if the shading/exclusions are still correct. Thanks.
There was a problem hiding this comment.
Yes, I have verified it. mvn dependency:tree outputs the same tree as before.
|
@walterddr fyi: this might be useful for the Flink planner Calcite upgrade as well. |
| <Resource name="planAfter"> | ||
| <![CDATA[ | ||
| Calc(select=[a, CAST(1984-07-12) AS b, CAST(14:34:24) AS c, CAST(1984-07-12 14:34:24) AS d], where=[AND(=(b, 1984-07-12), =(c, 14:34:24), =(d, 1984-07-12 14:34:24))]) | ||
| Calc(select=[a, CAST(1984-07-12) AS b, CAST(14:34:24) AS c, CAST(1984-07-12 14:34:24:TIMESTAMP(3)) AS d], where=[AND(=(b, 1984-07-12), =(c, 14:34:24), =(d, 1984-07-12 14:34:24:TIMESTAMP(3)))]) |
There was a problem hiding this comment.
I observe something weird upon comparing calcite 1.16/17/18/19 on this (not yet on 19) seems like the timestamp type representation is bouncing here and there a lot. do you think this is solely just the serialization (print string) changes or actually there's some changes to the underlying type? as this might be problematic.
There was a problem hiding this comment.
Hi @walterddr , what do you mean "timestamp type representation is bouncing here and there a lot"?
There was a problem hiding this comment.
@walterddr I think the reason is the modify of RexLiteral.toJavaString in apache/calcite#1002 , The Timestamp default precision is 3, I think it hasn't changed.
There was a problem hiding this comment.
looks like it. what I meant is that I've been changing the test cases string literals back and forth for some of the pass calcite upgrades. Seems like not an issue then.
|
LGTM, +1 to merge |
|
Thanks @godfreyhe . Hi @walterddr @twalthr @KurtYoung , I will merge this if you have no other comments. |
|
Merging... |
…in blink planner This closes apache#8234
What is the purpose of the change
Bump Calcite dependency to 1.19.0 in
flink-table-planner-blinkmodule.Brief change log
flink-table-planner-blink/pom.xmlFlinkRelOptTableto implement new interface method.Verifying this change
no new tests added.
Does this pull request potentially affect one of the following parts:
@Public(Evolving): (no)Documentation