[bug] partial OVERWRITE operation writes the wrong snapshot summary metrics

[kevinjqliu]

Apache Iceberg version

main (development)

Please describe the bug 🐞

Snapshot OVERWRITE operation can calculate the wrong summary fields when the table is partially updated.

update_snapshot_summaries assumes that all OVERWRITE operations are full table overwrite

iceberg-python/pyiceberg/table/update/snapshot.py

Line 239 in 322ebdd

truncate_full_table=self._operation == Operation.OVERWRITE,

iceberg-python/pyiceberg/table/snapshots.py

Lines 358 to 359 in 322ebdd

	if truncate_full_table and summary.operation == Operation.OVERWRITE and previous_summary is not None:
	summary = _truncate_table_summary(summary, previous_summary)

This is likely an oversight when we implemented partial write.

Thankfully the table/transaction's overwrite function is currently implemented as a delete+append.

The only place where OVERWRITE operation is used is during partial deletes.

iceberg-python/pyiceberg/table/__init__.py

Line 678 in 322ebdd

with self.update_snapshot(snapshot_properties=snapshot_properties).overwrite() as overwrite_snapshot:

Original thread apache/iceberg-go#356 (comment) (thanks @arnaudbriche and @zeroshade )

Partial overwrite reproduced in #1840

Willingness to contribute

• I can contribute a fix for this bug independently
• I would be willing to contribute a fix for this bug with guidance from the Iceberg community
• I cannot contribute a fix for this bug at this time