Pull a fair sample from a huge table for a quick check

My table has millions of rows, and I need a quick, honest look at it without pulling the whole thing.

Beginner to intermediateRead 5 minDoing: 20 to 30 minutes6 steps

What you'll have at the end

A few hundred rows spread across the whole table, not the same slice a plain first-page grab would return, ready for a fast check.

Before you start

You need

Read access to a table with millions of rows, including permission to run a full-table aggregate query such as an average.

Not covered

Building a stratified sample that guarantees coverage of a specific category, or turning this into a scheduled, repeating pipeline.

Leans on

6 steps

Where to go next

Compute percentiles in SQL for a table too big for a spreadsheet

when the quick check needs a median or a percentile instead of an average, since neither one can be estimated from a single fast aggregate the way an average can

Checked 26 Aug 2026

Part of the Descriptive Statistics cookbook