I don’t have the time (or the knowledge) to work on this right now, but I want to mention it in case anyone else might find it interesting: Deleting data usually requires log-based replication, but it should be possible to do the same using snowflake streams, because they are practically a log and marks rows as “insert” or “delete” in the correct order. I don’t know if it makes sense to add this as a feature to the existing tap or to create a tap-snowflake-streams, but it would be a fairly nice way of doing replication from snowflake.

cc @ken_payne

Agreed - but with a couple caveats. This would only apply for tables that are modified with a traditional row 'load' and 'delete' paradigm (basically oltp). However, dbt-managed tables are often drop-and-replace in each build, so those final-created tables from a dbt-managed process may not work as expected.

I agree with that caveat, in a dbt context only snapshots and incremental models make sense. When tables are replaced, any streams based on them are dropped so the E/L would fail unless the stream is recreated.