How to Learn SQL for Free, One Query at a Time

Short browser lessons take you from SELECT through joins, aggregates, and order of execution, then on to inserting rows and creating tables. Start here if you have never written a query and want to run one today.

An interactive book that teaches a mental model for building queries, using a real dataset about executions across four roughly 30-minute chapters. Read it after SQLBolt to learn why queries are structured as they are; the subject matter is heavy.

Jeff Atwood tests the popular Venn-diagram explanation of joins against real syntax, using two four-row tables to show what inner, full outer, and left outer joins return. Read it when join output stops making sense.

You get a missing crime-scene report and an SQLite database, and you write the SELECT, WHERE, and JOIN queries that identify the murderer. It is practice rather than a tutorial, so bring the basics from SQLBolt first.

Written for aspiring analysts who have used Excel but never coded, with an editor inside the lesson pages for running every example. ThoughtSpot, a BI vendor, publishes it, and it skips database setup, so developers need another source for that.

The official tutorial covers installing PostgreSQL, creating a database, then tables, joins, aggregates, views, foreign keys, transactions, and window functions. Take it once you want a real database on your own machine instead of a browser sandbox.

Seven mistakes that produce wrong or error-throwing aggregations, including confusing WHERE with HAVING, selecting columns missing from GROUP BY, and miscounting after a LEFT JOIN. The blog belongs to a course vendor, LearnSQL.

Defines a result set and a window from scratch, then builds up the OVER() clause, partitioning, and the categories of window functions with example queries. Read it before the manuals if the syntax looks alien.

Nine free CC0 datasets across 47 tables, with joins, nulls, and messy data, available as CSV or queryable in the browser with no account. Start with the 48-row movie rentals set, small enough to check answers by eye.

Loads the 120 Years of Olympics dataset from Kaggle into PostgreSQL, then works through 20 queries with problem statements and expected output. Use it to practise approaching real, unfamiliar tables instead of tidy tutorial data.

Explains first, second, and third normal form and why storing each fact in one place keeps data consistent. It also covers denormalization, so read it when you start designing tables rather than only querying them.

A working programmer's account of finishing Stanford's free nine-week Introduction to Databases online, covering relational design theory, SQL, indexes, and transactions. Read it to gauge the workload and see why a schedule with graded assignments got him to finish.

Markus Winand's free web edition of his paid book explains B-tree indexes and slow queries across Oracle, MySQL, PostgreSQL, and SQL Server; he also sells training. Read it once you write queries daily and want them fast.