◐ Shell
clean mode source ↗

docs: Fix Flink retrieval compatibility notes by franciscojavierarceo · Pull Request #6498 · feast-dev/feast

What this PR does / why we need it

Updates the Flink compute engine reference docs to match the merged runtime behavior.

The Flink engine now prefers native to_flink_table(table_env) retrieval jobs when available, but also supports the standard RetrievalJob.to_arrow() path by converting Arrow results into Flink tables. The docs still described Arrow/pandas-only retrieval jobs as unsupported, which is now stale.

Does this PR introduce a user-facing change

Yes.

The documentation now accurately describes which offline-store retrieval paths are supported by the Flink compute engine.

Testing

Docs-only change.

Related

Follow-up to the merged Flink engine work in #6476.