Inline Stored Procedures for hybrid tables in Snowflake

Inline Stored Procedures are designed specifically for operational workloads on hybrid tables. The entire procedure body executes as a single atomic unit pushed directly to the query processing layer, reducing per-statement overhead and delivering low latency for OLTP transactions.### Use casesImagine you need to update records in a transactional orders table and write audit logs in the same transaction with high concurrency and low latency.- Wrap single DML operations for abstraction or access control.- Execute multi-statement OLTP transactions atomically.- Schedule recurring maintenance by calling the procedure from a Snowflake task.### How to use```sql CREATE OR REPLACE INLINE PROCEDURE place_order( order_id NUMBER(18,0), customer_id NUMBER(18,0), product VARCHAR, quantity NUMBER(10,0) ) RETURNS VARCHAR LANGUAGE SQL AS $$ BEGIN ATOMIC INSERT INTO orders (order_id, customer_id, product, quantity, status) VALUES (:order_id, :customer_id, :product, :quantity, 'PENDING'); RETURN 'OK'; END; $$

CALL place_order(1001, 42, 'Snowboard', 2);

1. The command creates a stored procedure using an atomic block.
2. The statement inserts data into the hybrid table using the passed arguments.
3. The procedure returns a confirmation string indicating success.### Limitations and requirementsThe feature is generally available (GA). They operate exclusively on hybrid tables; statements accessing standard Snowflake tables are not supported. Explicit BEGIN, COMMIT, or ROLLBACK statements are not allowed inside the procedure body since each invocation runs as an atomic block.