fix: align multi-level indices before merge in to_dataframe()

[TaiSakuma]

Resolve #3694.

• Implements changes described in Data destruction during conversion to dataframe of nested arrays #3694 (comment)
• Modifies the expected value of an existing test in test_0080_flatpandas_multiindex_rows_and_columns.py
  • The new expected multi-level indices are more sensibly all tuples of integers, while the old ones contain floats and None.
• Adds a parametrized test to assert the issue is resolved.

[codecov]

Codecov Report

❌ Patch coverage is 87.50000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.71%. Comparing base (b749e49) to head (9746d3b).
⚠️ Report is 468 commits behind head on main.

Files with missing lines	Patch %	Lines
src/awkward/operations/ak_to_dataframe.py	87.50%	2 Missing ⚠️

Additional details and impacted files

Files with missing lines	Coverage Δ
src/awkward/operations/ak_to_dataframe.py	92.20% <87.50%> (+1.58%)	⬆️

... and 199 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

The documentation preview is ready to be viewed at http://preview.awkward-array.org.s3-website.us-east-1.amazonaws.com/PR3733