HAVING vs WHERE in SQL

In Excel, you'd reach for

fxFiltering the data vs filtering the pivot table totals

Both filter. The difference is when:

WHERE runs first, on individual rows, before anything is grouped. Use it for "only 2026 orders" or "only Europe".
HAVING runs after GROUP BY, on the totals. Use it for "only customers with 3+ orders" or "only months over $10k".

Writing WHERE COUNT(*) > 3 fails because when WHERE runs, nothing has been counted yet.

SQL order of execution: FROM, WHERE, GROUP BY, HAVING, SELECT, ORDER BY, LIMIT, with the Excel equivalent of each step — The order SQL runs a query in, with the Excel step each one is like. Download · Save to Pinterest

Syntax

SELECT col, SUM(x)
FROM table
WHERE row_condition
GROUP BY col
HAVING SUM(x) > 100;

Try itEdit the query, then Run

customer_id	orders	units
290	5	20
324	3	13
297	4	13
322	4	12
181	2	12
162	3	12
87	4	12
365	3	11
343	2	11
213	2	11

Showing 10 of 73 rows. Press Run to see them all.

Read it in order: take orders from April on (WHERE), bucket them per customer (GROUP BY), keep buckets with at least 2 orders (HAVING), sort (ORDER BY). In Excel that's filtering the source data, building a pivot, then applying a value filter to the pivot.

Your turn

Which countries have more than 20 customers who came from paid_search? Show country and customers.

This page on one card

HAVING vs WHERE in SQL: the Excel way (Filtering the data vs filtering the pivot table totals) and the SQL query, side by side — HAVING vs WHERE in SQL, on one card: the Excel way and the SQL. Download · Save to Pinterest

The practice data (7 tables)

customers: one row per customer
subscriptions: one row per paid plan
events: one row per thing someone did
orders: one row per order
products: a lookup table
campaigns: one row per channel per month
tickets: one row per support ticket

See every column and sample rows