Refactor Local Dynamic Filter test cases

[manupatteri]

Description

Dynamic Filter(DF) is a common functionality in multiple connectors. Since functionality is similar, similar tests are present in different connectors. This is part of an effort to extract such common tests and make it available at a common layer. This will allow newer connectors or connectors where DF is getting introduced to write tests quickly and correctly.

Since the original issue was created multiple PRs were already merged which takes care of different aspects of this.(For eg: #9193).

Hence I am re-raising a new PR which takes care of common tests for local dynamic filters only. AFAIK it is available mainly in Memory Connector and Hive Connector.

Is this change a fix, improvement, new feature, refactoring, or other?
This is refactoring of dynamic filter test cases. Fix for #5776

Is this a change to the core query engine, a connector, client library, or the SPI interfaces? (be specific)
It's a refactoring of the test cases in testing package.

How would you describe this change to a non-technical end user or system administrator?
There is no functional change. Just re-organizing tests for Dynamic Filtering to manage it better.

Related issues, pull requests, and links

Fixes #5776

Documentation

(x) No documentation is needed.
( ) Sufficient documentation is included in this PR.
( ) Documentation PR is available with #prnumber.
( ) Documentation issue #issuenumber is filed, and can be handled later.

Release notes

(x) No release notes entries required.
( ) Release notes entries required with the following suggested text:

# Section
* Fix some things. ({issue}`issuenumber`)

[hashhar]

Should these configs be contributed by AbstractLocalDynamicFilteringTest. That way all subclasses can use the constant defined there. I don't think these values would need to vary per connector.

[hashhar]

CI failure is related - new files have missing license header. you can add those by running ./mvnw license:format.

[raunaqmorarka]

I don't think the same test would make sense for both memory and hive connectors. In memory connector, MemoryPageSourceProvider implements dynamic filtering at row level for every value that it encounters. Memory connector is the only connector which performs dynamic filtering at a row level without requiring any special layout.
In hive connector, HivePageSourceProvider can perform dynamic filtering only if the probe side table is partitioned or bucketed or we're able to successfully prune row groups/stripes in parquet/orc files. So the tests here require appropriate layout for the probe table.
IMO we should just figure out what can be made common when the scenario of adding DF to some connector actually arises (e.g. we extracted BaseDynamicPartitionPruningTest when we added dynamic partition pruning to iceberg)

[manupatteri]

Ok. If that's the decision, then I can close this PR. Please let me know @sopel39 @hashhar. Thanks @raunaqmorarka.

[manupatteri]

CI failure is related - new files have missing license header. you can add those by running ./mvnw license:format.

Yeah. I realized this later. I think it's fixed now. Now I am getting something which looks a system error. I will just do a git push again to trigger a re-run.

[sopel39]

It's actually not about local dynamic filtering, but about how scheduler schedules stages (and it doesn't lock). So it's not local DF test. This maybe isn't the best place for it, but I couldn't find a better one

[sopel39]

I agree with @raunaqmorarka

[manupatteri]

Ok. Then I will go ahead and close this PR.
@sopel39 I see that Hive (non local DF) tests are already somewhat refactored. Do you want to close #5776 or does it need any more change?

[sopel39]

Do you want to close #5776 or does it need any more change?

Yes, I've closed #5576

[manupatteri]

Closing this PR as per discussion above.