SQL Playground

Create tables, insert rows and query them. A real SQLite database living in your browser tab, with the schema and results shown as you go.

← VizLearn

query.sql SQLite

-- A real SQLite database, running in this tab.
-- CREATE and INSERT persist until you reload or press Reset.

SELECT a.name,
       COUNT(*)        AS albums,
       SUM(al.sales)   AS total_sales
FROM   artists a
JOIN   albums  al ON al.artist_id = a.id
GROUP  BY a.name
HAVING SUM(al.sales) > 500000
ORDER  BY total_sales DESC;

Result

Tables

Try

CREATE TABLE your own, then insert into it — it stays until you reload.
A LEFT JOIN from artists: Nina Simone has a NULL country, so you can see how NULL behaves.
SELECT * FROM albums ORDER BY sales DESC LIMIT 3
A window function — RANK() OVER (ORDER BY sales DESC).

How it works

This is SQLite compiled to WebAssembly. The database lives in memory in this tab — nothing is uploaded, and a reload starts clean.

Ctrl/Cmd + Enter runs.

Which SQL this is

SQLite, compiled to WebAssembly and running in your browser. The database lives in memory: it is created when the page loads, it keeps your changes while you work, and Reset database puts the sample tables back exactly as they were.

SQLite is the most widely deployed database in the world and its dialect is close to standard SQL, so almost everything you write here transfers to PostgreSQL or MySQL. The differences worth knowing are small: SQLite is relaxed about types, storing whatever you give a column; it has no dedicated date type, using text or numbers instead; and || concatenates strings where MySQL would use CONCAT.

What is supported

SELECT with WHERE, GROUP BY, HAVING, ORDER BY and LIMIT.
All the joins — inner, left, cross, and self-joins.
Common table expressions, including recursive ones, and window functions such as ROW_NUMBER, RANK and SUM() OVER ().
INSERT, UPDATE, DELETE, CREATE TABLE and CREATE INDEX, so you can build your own schema and query it.
Transactions, and EXPLAIN QUERY PLAN for seeing how a query will be executed.

Errors you will see

no such column usually means a typo, or a column named in SELECT that is neither grouped nor aggregated. no such table means the name is wrong or the table was dropped — reset to bring the samples back. ambiguous column name means two joined tables share a column name, so it needs qualifying as table.column.

An empty result is not an error. It usually means the WHERE clause matched nothing, and the quickest way to find out is to remove conditions one at a time until rows appear.

Things worth trying

Group by two columns and watch the row count change, then add a column that is neither grouped nor aggregated and see what SQLite does with it.
Write the same question as a join, as a subquery, and as a CTE, and compare which one you would rather read in six months.
Put EXPLAIN QUERY PLAN in front of a query, add an index, and run it again to see the plan change.
Create your own table, insert a few rows, and break a constraint deliberately.

Learning SQL?

The Databases & SQL track works through the clauses in the order the engine actually applies them — which explains most of the errors people hit. Start with query execution order.