Deterministic Dictionary testing in CometFuzzTestSuite

[mbutrovich]

Is your feature request related to a problem or challenge?

#1697 exposed an issue in the random data generation: depending on RNG you can get Dictionary encoded values when written to Parquet, which can result in different paths being exercised.

Describe the solution you'd like

I tested this locally, but we could modify the generation to use a lower cardinality on some columns to force dictionary encoding.

Describe alternatives you've considered

No response

Additional context

No response

[mbutrovich]

Wrong repo, my bad :(