Rework some python tests of Parquet delta encodings

[etseidl]

Description

test_parquet.py currently takes around 55s to run on an RTXA6000 system. A large portion of that run time is in two tests of the Parquet DELTA_LENGTH_BYTE_ARRAY and DELTA_BYTE_ARRAY encodings. These tests are parameterized with varying row counts to test certain encoding edge cases, but the final two row counts (10,000, 50,000) are unnecessarily large to provide adequate test coverage. This PR reduces the number of row counts (some were redundant) and decreases the maximum row count to 1,000. This drops the execution time to just under 26s on the same system.

This PR also corrects an oversight from #15239. DELTA_BYTE_ARRAY encoding should have been added to the tests at that time.

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[vuule]

/ok to test

[vuule]

/merge