Understand COUNT(column) Versus COUNT(*)

`COUNT(column)` skips `NULL` values, while `COUNT(*)` counts every row that made it into the result set.

5/5 supported engines validation-green 1 example 2 scenarios

Docker-validated Not currently validation-green

Example 1

Compare total rows with only non-NULL timestamps

There are three invoice rows total, so COUNT() returns 3. Only two rows have a non-NULL paid_at, so COUNT(paid_at) returns 2. Use COUNT() when you want all rows, and COUNT(column) only when you intentionally want to ignore missing values.

Source table data Rows loaded before the example query runs.

Setup

CREATE TABLE invoices (id INT, paid_at VARCHAR(20));

INSERT INTO
  invoices (id, paid_at)
VALUES
  (1, '2026-04-01'),
  (2, NULL),
  (3, '2026-04-03');

Validated query Shared across supported engines.

SQL

SELECT
  COUNT(*) AS total_rows,
  COUNT(paid_at) AS rows_with_paid_at
FROM
  invoices;

Expected result Returned rows for the shared example.

total_rows	rows_with_paid_at
3	2

All engines treat `COUNT(column)` as a count of non-NULL values only.

Useful when

Where this command helps.

checking how many rows have a value in an optional column
avoiding undercounting when nullable columns are present

Explanation

What the command is doing.

A common counting mistake is assuming COUNT(column) behaves the same as COUNT(). It does not. COUNT(column) only counts rows where that specific column is not NULL, while COUNT() counts every row after filtering. This matters when optional fields, missing timestamps, or nullable foreign keys are involved.

Understand COUNT(column) Versus COUNT(*)

Compare total rows with only non-NULL timestamps

Where this command helps.

What the command is doing.

Keep moving through nearby tasks.