Each customer's largest order

Constraints

1. 1order_id must be the order that actually carries that customer's maximum.
2. 2Do not put order_id in the SELECT alongside a bare MAX(...) and hope it matches — nothing makes it.
3. 3Every customer with an order appears.
4. 4On this data no customer has two orders tied at their maximum, but write the query so a tie would return both rather than picking one arbitrarily.

Worked example

c1 placed three orders: 1001 for 120, 1003 for 45 and 1008 for 300. The answer is order 1008 — and 300 on its own is not the answer, because the question asks which order carried it. SELECT customer_id, order_id, MAX(order_total) ... GROUP BY customer_id looks like it works. It reports 300 for c1, but the order_id printed beside it is whichever row the engine happened to keep, which may well be 1001. Most engines reject that query outright; SQLite accepts it and quietly returns something plausible, which is worse.

What this tests

That an aggregate collapses a group to a value, so identifying the row behind that value takes a second step rather than an extra column in the SELECT.