Full Source Table Name Support
Add the full source table name as a column, and optionally to the primary key, so sharded tables can fan into one destination table without key collisions.
When data is sharded across multiple upstream tables, primary keys are often only unique within each table-not across all of them. Until now, that meant fanning data into a single destination table wasn't possible without risking collisions.
You can now include the complete source table name ({{database}}.{{schema}}.{{table}}) as a column in your replication stream. This provides additional metadata about where each row originated. You can also choose to make this new column part of the primary key - giving you a surrogate key that ensures uniqueness across shards.
Why this matters:
- Replicate sharded datasets into a single destination table without key conflicts
- Preserve visibility into the exact origin of each record
- Eliminate the need for complex workarounds or manual deduplication
- Unlock new fan-in use cases for teams consolidating data from many sources