Append-only option for Spark incremental append read that throws #13179

Feature Request / Improvement

Spark's incremental append reading is useful for append-only incremental workflows.

It's documented as only reading snapshots with "append" snapshot types - it ignores any snapshots in the range that don't have this type, e.g. DELETE snapshots.

I think this is sensible default behaviour, but I'm wondering what the community thinks about allowing an append-only option to be set that throws if there are non-append snapshots in the range, to mitigate data integrity of a downstream table being compromised for strictly append-only workflows. (It would be unset by default)

Query engine

None

Willingness to contribute

• I can contribute this improvement/feature independently
• I would be willing to contribute this improvement/feature with guidance from the Iceberg community
• I cannot contribute this improvement/feature at this time