DISTINCT vs GROUP BY: which one should you use?

SELECT DISTINCT col FROM t and SELECT col FROM t GROUP BY col return the same rows. So why prefer one?

DISTINCT is for de-duping projection

Use DISTINCT when you have no aggregates and you genuinely want unique tuples of the projected columns. It is shorter and clearer for that case.

GROUP BY is for aggregation

The moment you add COUNT, SUM, or any aggregate, you must use GROUP BY. DISTINCT cannot express 'group these and aggregate those' in one step.

Why DISTINCT often hides bugs

DISTINCT silently masks fan-out joins. The row count looks reasonable because duplicates were collapsed — but the bug is still there, ready to surface when someone adds an aggregate later.

-- Looks fine, but if order_items fans out, DISTINCT hides it.
SELECT DISTINCT o.order_id, o.customer_id
FROM orders o JOIN order_items i ON i.order_id = o.order_id;