Add test coverage for grouping on dictionary encoded columns

[alamb]

Is your feature request related to a problem or challenge?

At InfluxData we use Dictionary(Int32, Utf8) heavily for certain columns of our queries.

This particular type of data is not well represented in the DataFusion codebase, which leads to issues such as #8738

In addition to not being well covered by functional tests, we also don't have particularly good coverage performance wise, especially for some of our important cases like grouping performance. This leads to dueling speculations on better/worse ways to run such queries, such as #7647 but now good way to evaluate them

Describe the solution you'd like

I would like to ensure we have adequate test coverage

1. Functional tests so we don't hit regressions accidentally again
2. Performance tests (e.g. so we could evaluate the changes contemplated by Materialize Dictionaries in Group Keys #7647

Describe alternatives you've considered

I added some functional sqllogictests in

• https://github.com/apache/arrow-datafusion/blob/main/datafusion/sqllogictest/test_files/dictionary.slt

However, those are not reading from Parquet files (as I could not figure out how to make a parquet file have the same schema I wanted)

Additional context

No response