Every SQL keyword belongs to one of four families. Learn them once, never confuse them again.

SQL Command Categories

There are dozens of SQL keywords. Memorizing them is pointless. Grouping them into four families makes them stick.

The four families

Family	What it does	Keywords you'll see
DDL — Data Definition Language	Defines the SHAPE of your data (creates / changes / removes tables)	`CREATE`, `ALTER`, `DROP`, `TRUNCATE`, `RENAME`
DML — Data Manipulation Language	Changes the CONTENT of your data (rows in/out)	`INSERT`, `UPDATE`, `DELETE`, `MERGE`
DCL — Data Control Language	Manages PERMISSIONS (who can do what)	`GRANT`, `REVOKE`
DQL — Data Query Language	Reads data (the most common one by far)	`SELECT`

Some textbooks also list TCL — Transaction Control Language (COMMIT, ROLLBACK, SAVEPOINT) as a fifth family. We'll cover transactions in Module 9.

Why this grouping matters

It tells you what the command affects and how dangerous it is:

DDL changes the structure — affects every row in a table at once. Hard to reverse. Big risk.
DML changes rows — affects specific rows. Reversible via transactions before commit.
DCL changes who-can-do-what — affects security. Production teams audit this.
DQL just reads — zero risk. The only family safe to run on production without a second pair of eyes.

When you see a query, ask: which family? If it's DDL, slow down. If it's DQL, just run it.

Each family in one example

DDL — define structure

CREATE TABLE customers (
  id INT PRIMARY KEY,
  name VARCHAR(100) NOT NULL,
  email VARCHAR(255) UNIQUE
);

This creates a new table. Until you run an INSERT, the table has zero rows.

DML — change content

INSERT INTO customers (id, name, email)
VALUES (1, 'Anita Rao', 'anita@example.com');

This adds one row to the existing table.

DCL — control access

GRANT SELECT ON customers TO analytics_user;

This gives the analytics_user role permission to read from the customers table, but not write.

DQL — read

SELECT name, email FROM customers WHERE id = 1;

This reads one row matching id = 1. Returns Anita Rao / anita@example.com.

A common interview trap

"What's the difference between DELETE and TRUNCATE?"

Both empty out a table. The trick:

`DELETE`	`TRUNCATE`
DML — manipulates rows	DDL — redefines table
Row by row	Whole table at once
Logs each row deletion	Doesn't log individual rows (faster)
Can have a `WHERE` clause	All-or-nothing
Can be rolled back inside a transaction	Can't always be rolled back
Triggers fire	Triggers do NOT fire

DELETE is surgical. TRUNCATE is a wrecking ball. Same end state, very different mechanics.

Cheat sheet to bookmark

DDL → defines structure         CREATE / ALTER / DROP / TRUNCATE
DML → changes content           INSERT / UPDATE / DELETE
DCL → manages permissions       GRANT / REVOKE
DQL → reads data                SELECT
TCL → transaction control       COMMIT / ROLLBACK / SAVEPOINT

Print this and tape it to your monitor for week one of any SQL job.

Module 1 complete 🎉

You now know:

✅ What SQL is and why it's worth learning
✅ Real-world scenarios where SQL is the answer
✅ What an RDBMS is and the major players
✅ Tables, rows, columns, fields, NULL
✅ The four families of SQL commands

In Module 2: SQL Databases, we'll do a quick tour of the major RDBMS systems — MySQL, PostgreSQL, SQLite, MS SQL Server, Oracle — and help you pick which one to install for hands-on practice.

SQL Command Categories (DDL/DML/DCL/DQL)