[FLINK-36662][table-planner] Port CALCITE-6317 to fix the incorrectly constant pullup when group keys in aggregate are NULL

[lincoln-lil]

What is the purpose of the change

This pr port the bugfix of CALCITE-6317 to fix the incorrectly constant pullup when group keys in aggregate are NULL
Before upgraded corresponding calcite version(1.37), we can have the fix that copy the related RelMdPredicates to flink and remove it after calcite version upgrading done.

Brief change log

Copy the RelMdPredicates based on 1.32.0 and add the fix of CALCITE-6317

Verifying this change

Add new case for GroupingSetsTest

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): (no)
• The public API, i.e., is any changed class annotated with @public(Evolving): (no)
• The serializers: (no )
• The runtime per-record code paths (performance sensitive): (no)
• Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (no)
• The S3 file system connector: (no)

Documentation

• Does this pull request introduce a new feature? (no)

[flinkbot]

CI report:

• e341c8f Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[snuyanzin]

Can you please mark the changed lines with

// ------ FLINK MODIFICATION BEGIN
...
// ------ FLINK MODIFICATION END

and put numbers of lines in class comment as it is done for other Calcite's classes

it will significantly help during upgrade (before we reach 1.37)

[lincoln-lil]

@snuyanzin thank you for reviewing this so quickly! (My original plan was to use the night hours (utc+8) to do a full assessment of the impact of the existing tests)
I've now added a description of the scope for the changes.

[snuyanzin]

LGTM
thanks for addressing feedback

[lincoln-lil]

LGTM thanks for addressing feedback

@snuyanzin Thank you! Merging...