Interesting article to use Databricks SQL for data integration. Just one question, why would you write the INSERT in a separate statement when doing the MERGE? This could be extremely dangerous and introduce data inconsistency since it's not treated as an atomic transaction.