[ShivamSharma43]

Rationale for this change

When a snapshot is removed (RemoveSnapshotsUpdate, which expire_snapshots uses), pyiceberg removes the snapshot's statistics files but keeps its partition statistics files. After expiring a snapshot, the table metadata still has partition-statistics entries pointing at a snapshot that is gone.

Iceberg Java removes both in TableMetadata.Builder.rewriteSnapshotsInternal:

removeStatistics(snapshotId);
removePartitionStatistics(snapshotId);

This adds the same removal for partition statistics, next to the existing one for statistics. A REST catalog server applies the update itself, so you mostly see this with catalogs where pyiceberg writes the metadata (SQL, Glue, Hive, DynamoDB). The test for #4201 works around the same gap by patching commit_table to drop these entries.

I didn't find an existing issue for this.

Are these changes tested?

Yes. I added test_update_remove_snapshots_with_partition_statistics to tests/table/test_expire_snapshots.py, next to the existing statistics test. It fails on main (the partition statistics entry for the removed snapshot is still there) and passes with this change. prek run passes on the changed files.

Are there any user-facing changes?

After expire_snapshots, table metadata no longer lists partition statistics for snapshots that were removed.

AI disclosure

I used Claude Code (AI) to find this gap by comparing pyiceberg with Iceberg Java, and to write the fix and the test. I reviewed the change. One thing I'm not sure about: I kept the removal in the RemoveSnapshotsUpdate handler, applied as internal RemovePartitionStatisticsUpdates the same way the statistics removal already works. I did not add a matching cleanup of the partition statistics files on storage. That belongs with the file-cleanup work in #4201.

🤖 Generated with Claude Code

https://claude.ai/code/session_012XPuJeHi7A5CeHSKHjHmwx

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.