Every night, a MERGE synchronizes catalog.sales.silver_orders with the source extract, updating matched orders and inserting new ones: ```sql MERGE INTO catalog.sales.silver_orders AS t USING catalog.sales.source_orders AS s ON t.order_id = s.order_id WHEN MATCHED THEN UPDATE SET * WHEN NOT MATCHED THEN INSERT * ``` Orders are sometimes deleted in the source, and those deletions must also disappear from silver_orders, in the same atomic statement. Which clause adds this behavior?

1. A. WHEN NOT MATCHED BY SOURCE THEN DELETE — it applies to the target rows whose keys have no counterpart in the source table ✓ A
2. B. WHEN NOT MATCHED THEN DELETE — that clause already covers the target rows that are absent from the source 💬 B
3. C. None: every MERGE clause can only act on rows of the source, so unmatched target rows are out of the statement's reach 💬 C
4. D. Deletion is outside MERGE's scope, so schedule a separate DELETE ... WHERE order_id NOT IN (SELECT ...) after every merge runs 💬 D

Reveal answer

Correct answer: WHEN NOT MATCHED BY SOURCE THEN DELETE — it applies to the target rows whose keys have no counterpart in the source table

Reference: https://docs.databricks.com/aws/en/delta/merge