Skip to content

Conversation

@010Soham
Copy link
Contributor

@010Soham 010Soham commented Dec 19, 2025

What does this change do?

  • Build the inspect.partitions schema using only the partition specs present in the selected snapshot’s manifests, so newer partition fields don’t appear as None for older snapshots.
  • Normalize partition comparisons in integration tests to ignore trailing None fields from dropped partition columns.

Why is this needed?

  • Partition evolution should show the partition shape that actually existed when the data was written. Mixing all specs produced misleading None fields for older partitions.

How was this tested?

  • make lint
  • uv run python -m pytest tests/io/test_pyarrow.py -k "partition_evolution" -v
  • CI: full integration suite (GitHub Actions)

Closes #1120

Copy link
Contributor

@kevinjqliu kevinjqliu left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! Thanks!

@kevinjqliu kevinjqliu merged commit b0a7878 into apache:main Dec 24, 2025
8 checks passed
@kevinjqliu
Copy link
Contributor

Thanks for the PR @010Soham

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[bug] table.inspect.partitions() does not respect partition evolution

2 participants