Details
-
New Feature
-
Status: Open
-
P3
-
Resolution: Unresolved
-
None
-
None
-
None
Description
Databricks Delta is an open source storage layer on top of different filesystems. The current implementation of Delta is strongly coupled with Spark so we cannot rely on it because it would break Beam portability.
However now there is an open specification for Delta's protocol.
https://github.com/delta-io/delta/blob/master/PROTOCOL.md
Another possible approach could be to investigate how if Beam could use a manifest based approach like Presto does:
https://docs.databricks.com/delta/presto-integration.html