Why should you care?

Standardized “place order” logic shared by API and batch job avoids copy-paste bugs across Node and Python services.

See it live — copy this example

Run in MySQL Workbench or the mysql CLI.

DELIMITER //
CREATE PROCEDURE sp_dataflow_place_order(
  IN p_customer_id INT UNSIGNED,
  IN p_total DECIMAL(12,2),
  OUT p_order_id INT UNSIGNED
)
BEGIN
  INSERT INTO orders (customer_id, order_ref, total_inr)
  VALUES (p_customer_id, CONCAT('DF-', UUID()), p_total);
  SET p_order_id = LAST_INSERT_ID();
END //
DELIMITER ;
CALL sp_dataflow_place_order(1, 250.00, @oid);
SELECT @oid;

What happened?

• Procedure inserts order with generated ref and returns new id via OUT param.
• CALL from app or Workbench.
• Logic lives in MySQL — deploy with migrations.

Practice next

1. Run CREATE PROCEDURE in Workbench (allow multi-statements).
2. CALL with sample customer_id.
3. SELECT @oid and verify orders row.
4. Add DECLARE handler for duplicate order_ref.
5. Call from mysql CLI: CALL sp_dataflow_place_order(1,100,@oid);

Remember

Procedures bundle SQL with parameters. OUT/INOUT return values to caller. Version-control like application code.

DataFlow order API

Node uses CALL sp_dataflow_place_order(?,?,@oid) for consistent inserts.

Outcome: One definition; all clients behave identically.