Level 1 · Find refunded orders

WHERE keeps only the rows for which a condition is true. It goes after FROM:

SELECT order_id, customer_id, order_date, channel
FROM orders
WHERE status = 'refunded';

You can filter on a column you do not select. status decides which rows survive, but it does not have to appear in the output.

Comparison operators

operator	meaning
=	equal
<> or !=	not equal (<> is the SQL standard; != works almost everywhere)
<, <=, >, >=	less / greater than

Numbers are written bare (shipping_fee > 10), text and dates go in single quotes (channel = 'app', signup_date >= '2025-01-01'). Double quotes mean something else in Postgres and DuckDB: they quote an identifier such as a column name. WHERE status = "refunded" looks for a column called refunded and fails. MySQL accepts double quotes for strings by default, which is a habit worth unlearning.

Exact matches are exact

'refunded', 'Refunded' and 'refunded ' (with a trailing space) are three different values in Postgres, DuckDB and BigQuery. MySQL compares text case-insensitively by default, which hides messy data until the query moves to another database. Before filtering a text column, run SELECT DISTINCT status FROM orders; and copy the exact values you see.

How the database reads your query

You write SELECT ... FROM ... WHERE, but the database logically runs FROM first, then WHERE, then SELECT. That is why you cannot use a column alias from the SELECT list inside WHERE in most databases:

-- fails in Postgres, MySQL and BigQuery
SELECT order_id, shipping_fee * 2 AS double_fee
FROM orders
WHERE double_fee > 20;

Repeat the expression in WHERE instead. DuckDB is lenient and allows the alias, but portable SQL does not rely on it. This “logical order” idea comes back with GROUP BY and HAVING in chapter 4, where it is a favorite interview question.

Common mistakes

• Filtering on the wrong value spelling, then reporting “zero refunds”. If a filter returns nothing, check the distinct values before concluding.
• Filtering on a number stored as text. WHERE shipping_fee = '6.95' works in DuckDB and Postgres because the literal is converted, but comparing text columns with numbers can fail or compare alphabetically in other databases. Match the column’s type.
• Using = for a list of values. status = 'refunded', 'cancelled' is not valid; you need IN, which is next.