developer-roadmap/src/data/courses/sql/chapters/data-definition-language/lessons/primary-key-constraint.md

title	description	order	type
Primary Key Constraint	Learn about primary keys and how to use them effectively.	160	lesson-challenge

In our previous lessons, we learned about various constraints that help maintain data integrity. In this lesson we will learn about primary keys and how to use them effectively.

Primary Keys

A primary key is a column (or a combination of columns) that uniquely identifies each row in a table. Think of it as a unique ID that helps us reference specific rows in our table.

Here's an example of a table with a primary key:

-- Creating a table with a primary key
CREATE TABLE books (
  id INTEGER PRIMARY KEY,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13),
  price DECIMAL(10, 2)
);

Alternatively, there are two other syntaxes for creating primary keys:

-- Define primary key in the constratins section
CREATE TABLE books (
  id INTEGER,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13),
  price DECIMAL(10, 2),

  PRIMARY KEY (id)
);

-- You can also name the constraint so that database
-- shows it in the error messages
CREATE TABLE books (
  id INTEGER,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13),
  price DECIMAL(10, 2),

  CONSTRAINT books_identifier PRIMARY KEY (id)
  --         -------^--------
  --      Name of the constraint
);

In this example, the id column is our primary key. Each book in the table must have a unique id value. For example, if we try to insert a book without and id, it will fail:

-- Error: Null value in column 'id' violates not-null constraint
-- Primary key columns cannot be null
INSERT INTO books (title, isbn, price)
VALUES ('Book 1', '1234567890123', 19.99);

Similarly, if we try to insert a book with an existing id, it will fail:

-- Error: duplicate key value violates unique constraint "books_pkey"
INSERT INTO books (id, title, isbn, price)
VALUES (1, 'Book 1', 'xxxxxxxxxxxxx', 9.99),
       (1, 'Book 2', 'yyyyyyyyyyyyy', 20.99);

Why are Primary Keys Important?

In our previous lessons, we haven't been creating primary keys as I was postponing this lesson until now. But it is highly recommended to define primary keys for each table, because:

• They help uniquely identify each row in a table.
• We can prevent duplicate rows from being inserted.
• They can be used to create relationships between tables.
• Primary keys are indexed by default, resulting in improved query performance.

We will learn about indexes in the future lessons.

Auto-generating Primary Keys

Most databases provide a way to automatically generate primary key values. It is not a part of the SQL standard, but many databases have their own syntax for it e.g.

Database	Syntax
PostgreSQL	SERIAL
MySQL	AUTO_INCREMENT
SQLite	AUTOINCREMENT
SQL Server	IDENTITY
Oracle	SEQUENCE

The example below shows how to create a table with a primary key that is auto-generated in PostgreSQL:

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

-- Now we can insert without specifying id
INSERT INTO authors (name, email)
VALUES ('J.K. Rowling', 'jk@example.com'),
       ('Stephen King', 'stephen@example.com'),
       ('Agatha Christie', 'agatha@example.com');

Now if we fetch all the rows from the authors table:

SELECT * FROM authors;

We will see the auto-generated id values for each row:

id	name	email
1	J.K. Rowling	jk@example.com
2	Stephen King	stephen@example.com
3	Agatha Christie	agatha@example.com

The database will automatically assign and increment the id value for each new row.

Composite Primary Keys

Sometimes, you might need multiple columns to uniquely identify a row. This is called a composite primary key:

CREATE TABLE book_editions (
  book_id INTEGER,
  edition_number INTEGER,
  publication_year INTEGER,

  -- Creating a composite primary key
  PRIMARY KEY (book_id, edition_number)
);

In this case, a specific edition of a book is uniquely identified by both its book_id and edition_number. We can't insert a row with the same book_id and edition_number twice.

For example, if we try to insert a row with the same book_id and edition_number:

-- Error: duplicate key value violates unique constraint "book_editions_pkey"
INSERT INTO book_editions (book_id, edition_number, publication_year)
VALUES (1, 1, 2020),
       (1, 1, 2021);

If we try to insert a row with a different book_id:

-- This will work because the combination of book_id and edition_number is unique
INSERT INTO book_editions (book_id, edition_number, publication_year)
VALUES (1, 1, 2020),
       (2, 1, 2021);

While we are on the topic of primary keys and uniqueness, let's also look at the UNIQUE constraint which can be used to enforce uniqueness on any column or a combination of columns.

UNIQUE Constraint

While primary keys enforce uniqueness, sometimes you need other columns to be unique as well. The UNIQUE constraint ensures no duplicate values can exist in a column:

CREATE TABLE books (
  id INTEGER PRIMARY KEY,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13) UNIQUE, -- This column must be unique
  price DECIMAL(10, 2),
);

Again, just like any other constraint, here is the alternative syntax:

-- Define in constraints section
CREATE TABLE books (
  id INTEGER PRIMARY KEY,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13),
  price DECIMAL(10, 2),

  CONSTRAINT books_isbn_unique UNIQUE (isbn)
);

-- Or without naming the constraint
CREATE TABLE books (
  id INTEGER PRIMARY KEY,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13),
  price DECIMAL(10, 2),

  UNIQUE (isbn)
);

In this example, isbn must be unique across all rows. For example, if we try to insert a row with the same isbn:

-- Error: duplicate key value violates unique constraint "books_isbn_unique"
INSERT INTO books (id, title, isbn, price)
VALUES (1, 'Book 1', '1234567890123', 19.99),
       (2, 'Book 2', '1234567890123', 20.99);

The difference between UNIQUE and PRIMARY KEY is that a table can have multiple UNIQUE columns but only one primary key, UNIQUE columns can also contain NULL values (unless specified otherwise), while primary keys cannot contain NULL values.

NULL Considered UNIQUE

By default, UNIQUE constraint allows NULL values. You can change that behavior by adding NULLS NOT DISTINCT to the constraint i.e.

CREATE TABLE books (
  id INTEGER PRIMARY KEY,
  title VARCHAR(255) NOT NULL,
  isbn VARCHAR(13) UNIQUE NULLS NOT DISTINCT,
  --                      ---------^--------
  price DECIMAL(10, 2),
);

In this case, NULL values are not considered unique and you will get an error if you try to have multiple rows with NULL values in the isbn column e.g.

-- Error: We are trying to insert two rows with NULL values in the isbn column
INSERT INTO books (id, title, isbn, price)
VALUES (1, 'Book 1', "XYZ", 19.99),
       (2, 'Book 2', NULL, 20.99),
       (3, 'Book 3', NULL, 21.99);