SELECT: find the needed rows without changing them

Why this matters

A database is useful because it returns exactly what is needed: recent spending, purchases above a threshold, or one month’s events.

Image. FROM puts the receipt box on the table; WHERE keeps matching receipts; SELECT covers unused fields; ORDER BY arranges them; LIMIT takes the first few.

See the whole thing first

SELECT happened_on, amount_tiyn / 100.0 AS amount_kzt, note
FROM transactions
WHERE amount_tiyn >= 1000000
ORDER BY happened_on DESC, id DESC
LIMIT 5;

Explanation

The logical order begins with FROM and WHERE even though SELECT is written first. SELECT * is fine for exploration, but application contracts should name columns. Without ORDER BY row order is undefined; add id as a tie-breaker. LIKE rules, especially case handling, differ across systems.

Lesson map

Say it in your own words

1. Why is SELECT written first but not evaluated first?
2. What resolves equal dates?
3. When is SELECT * risky?

Exercise

Return September’s three largest transactions with date, tenge amount, and note; break ties by newest row.

Where this fits in the project

This query pattern powers transaction history and previews rows before a bulk change.

Answers

SELECT happened_on, amount_tiyn / 100.0 AS amount_kzt, note
FROM transactions
WHERE happened_on >= '2026-09-01' AND happened_on < '2026-10-01'
ORDER BY amount_tiyn DESC, happened_on DESC, id DESC
LIMIT 3;

Sources

• SQLite: SELECT