Pandas to_sql to INSERT / staging table

if_exists="replace" drops the table. Do not point that at something named Customers.

INSERT INTO OrdersStage (order_id, customer_id, total_amount)
SELECT order_id, customer_id, total_amount FROM #incoming;

A million rows: bcp, BULK INSERT, or SqlBulkCopy. to_sql method="multi" is still the app talking too much.

Canonical translation

INSERT INTO OrdersStage (order_id, customer_id, total_amount)
SELECT order_id, customer_id, total_amount FROM #incoming;