COALESCE in SQL is the single most-asked NULL-handling primitive at every data-engineering interview that touches production analytics. The mental model: COALESCE(expr1, expr2, ..., exprN) returns the first argument that is not NULL , evaluated left-to-right; if every argument is NULL , the result is NULL . That one rule covers a vast surface — fallback chains for "best available column," default values to neutralize LEFT JOIN misses, sentinel labels like 'NONE' for missing dimensions, and zero-substitution before downstream math. Four sub-primitives carry the loop: left-to-right evaluation with short-circuit semantics, the COALESCE(left_join_col, default) pattern that turns outer-join NULL s into reportable defaults, the dialect-portability matrix versus CASE / ISNULL / NVL / IFNULL , and the pitfall set around NULL ≠ 0 semantics, type coercion, empty strings, and the COALESCE(NULLIF(col, ''), 'default') composition.

This guide walks four concept clusters end-to-end, each with a detailed topic explanation…