WHERE

Filter rows before grouping, and understand how NULL and join filters change the result.

Syntax

SELECT ...
FROM table
WHERE predicate [AND | OR predicate ...];

WHERE removes rows before GROUP BY, aggregates, and window functions run. A row is kept only when the predicate evaluates to TRUE — a predicate that evaluates to FALSE or UNKNOWN drops the row.

Example

SELECT
  customer_id,
  amount
FROM orders
WHERE status = 'paid'
  AND amount > 0;

NULL makes predicates UNKNOWN

WHERE cancelled_at = NULL      -- never matches
WHERE cancelled_at IS NULL     -- correct

Any comparison with NULL returns UNKNOWN, so WHERE col <> 'x' silently drops rows where col is NULL. Add OR col IS NULL when those rows should stay.

LEFT JOIN filters belong in ON

-- turns the LEFT JOIN into an INNER JOIN
FROM customers c
LEFT JOIN orders o ON c.customer_id = o.customer_id
WHERE o.status = 'paid';

-- keeps unmatched customers
FROM customers c
LEFT JOIN orders o
  ON c.customer_id = o.customer_id
  AND o.status = 'paid';

A predicate on the right-side table in WHERE removes the NULL-filled unmatched rows, collapsing the outer join. Move the condition into ON when you need to keep them.

Safety check before shipping

decide whether NULL rows should pass or be excluded for each predicate
confirm right-side filters on outer joins are in ON, not WHERE, unless the collapse is intended
check row counts before and after adding a filter