Serial and identity columns in PostgreSQL - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a serial column in PostgreSQL?

A serial column is a special integer column that automatically increments its value for each new row. It uses a sequence behind the scenes to generate unique numbers.

intermediate

How does an identity column differ from a serial column?

An identity column is part of the SQL standard and provides automatic numbering like serial, but it is more flexible and preferred in modern PostgreSQL versions. It can be defined as GENERATED ALWAYS or GENERATED BY DEFAULT.

beginner

What SQL syntax creates a serial column?

Example:
CREATE TABLE users (id serial PRIMARY KEY, name text);
This creates an integer column id that auto-increments with each new row.

intermediate

Write the SQL to create an identity column named id that always generates a new value.

Example:
CREATE TABLE users (id int GENERATED ALWAYS AS IDENTITY PRIMARY KEY, name text);
This creates an id column that automatically generates unique values and does not allow manual inserts.

intermediate

Why might you choose identity columns over serial columns?

Because identity columns follow the SQL standard, offer more control, and avoid creating implicit sequences, making your database design cleaner and more portable.

Which PostgreSQL column type automatically creates a sequence and increments values?

A serial

B varchar

C text

D boolean

The serial type automatically creates a sequence and increments integer values for new rows.

What keyword is used to define an identity column that always generates a value?

A GENERATED BY DEFAULT

B GENERATED ALWAYS

C AUTO_INCREMENT

D SERIAL

GENERATED ALWAYS AS IDENTITY means the database always generates the value automatically.

Which of these is a benefit of using identity columns over serial columns?

A They are part of the SQL standard

B They allow text values

C They do not auto-increment

D They require manual sequence creation

Identity columns follow the SQL standard, making them more portable and standardized.

What happens if you insert a row without specifying a value for a serial column?

A You must provide a value manually

B The insert fails with an error

C The column is set to NULL

D The database generates the next number automatically

Serial columns auto-increment, so the database fills in the next number automatically.

Which statement is true about GENERATED BY DEFAULT AS IDENTITY?

A It is the same as <code>serial</code>

B The database always generates the value and disallows manual inserts

C You can insert your own value or let the database generate one

D It disables auto-increment

GENERATED BY DEFAULT allows manual insertion of values but generates one if none is provided.

Explain the difference between serial and identity columns in PostgreSQL.

Think about how the numbering is generated and standards.

Describe how to create a table with an auto-incrementing primary key using identity columns.

Focus on the syntax for identity columns.

Practice

1.

What is the main purpose of using SERIAL or IDENTITY columns in PostgreSQL?

easy

A. To automatically generate unique numbers for new rows

B. To store large text data efficiently

C. To create foreign key relationships

D. To improve query performance by indexing

Solution

1. Step 1: Understand the role of SERIAL and IDENTITY

   Both SERIAL and IDENTITY columns are designed to automatically generate unique numeric values for new rows in a table.

2. Step 2: Identify the main purpose

   This automatic numbering helps avoid manual entry and ensures unique identifiers for each row.

3. Final Answer:

   To automatically generate unique numbers for new rows -> Option A

4. Quick Check:

   Serial and identity columns = auto unique numbers [OK]

Hint: They auto-generate unique IDs for new rows [OK]

Common Mistakes:

• Thinking SERIAL stores text data
• Confusing SERIAL with indexing
• Assuming SERIAL creates foreign keys

2.

Which of the following is the correct syntax to create a table with an IDENTITY column named id in PostgreSQL?

CREATE TABLE users (
  id ???,
  name TEXT
);

easy

A. INT GENERATED AS IDENTITY

B. SERIAL PRIMARY KEY

C. INTEGER GENERATED ALWAYS AS IDENTITY PRIMARY KEY

D. INTEGER AUTO_INCREMENT PRIMARY KEY

Solution

1. Step 1: Recall IDENTITY syntax in PostgreSQL

   The correct syntax uses GENERATED ALWAYS AS IDENTITY for identity columns.

2. Step 2: Match the option with correct syntax

   INTEGER GENERATED ALWAYS AS IDENTITY PRIMARY KEY uses INTEGER GENERATED ALWAYS AS IDENTITY PRIMARY KEY, which is valid PostgreSQL syntax.

3. Final Answer:

   INTEGER GENERATED ALWAYS AS IDENTITY PRIMARY KEY -> Option C

4. Quick Check:

   IDENTITY syntax = GENERATED ALWAYS AS IDENTITY [OK]

Hint: IDENTITY uses GENERATED ALWAYS AS IDENTITY syntax [OK]

Common Mistakes:

• Using AUTO_INCREMENT (MySQL syntax)
• Confusing SERIAL with IDENTITY syntax
• Omitting GENERATED ALWAYS or BY DEFAULT

3.

Given the table creation:

CREATE TABLE products (
  product_id SERIAL PRIMARY KEY,
  name TEXT
);

INSERT INTO products (name) VALUES ('Pen'), ('Pencil'), ('Eraser');

SELECT product_id FROM products ORDER BY product_id;

What will be the output of the SELECT query?

medium

A. [3, 2, 1]

B. [0, 1, 2]

C. [NULL, NULL, NULL]

D. [1, 2, 3]

Solution

1. Step 1: Understand SERIAL behavior on insert

   The SERIAL column auto-increments starting from 1 for each inserted row.

2. Step 2: Check inserted rows and ordering

   Three rows inserted get product_id values 1, 2, and 3 respectively. Ordering by product_id returns them in ascending order.

3. Final Answer:

   [1, 2, 3] -> Option D

4. Quick Check:

   SERIAL starts at 1 and increments [OK]

Hint: SERIAL starts numbering at 1 by default [OK]

Common Mistakes:

• Assuming numbering starts at 0
• Expecting NULL values for SERIAL
• Thinking order is descending by default

4.

Consider this table creation:

CREATE TABLE orders (
  order_id INT GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY,
  product TEXT
);

Which of the following statements will cause an error when inserting data?

medium

A. INSERT INTO orders (order_id, product) VALUES (5, 'Pen'), (5, 'Pencil');

B. INSERT INTO orders (order_id, product) VALUES (5, 'Pen');

C. INSERT INTO orders (product) VALUES ('Book');

D. INSERT INTO orders (order_id, product) VALUES (DEFAULT, 'Notebook');

Solution

1. Step 1: Understand GENERATED BY DEFAULT AS IDENTITY behavior

   This allows manual insertion of order_id or automatic generation if omitted.

2. Step 2: Analyze each insert statement

   INSERT INTO orders (order_id, product) VALUES (5, 'Pen'), (5, 'Pencil'); tries to insert two rows with the same order_id = 5, violating the primary key uniqueness constraint, causing an error.

3. Final Answer:

   INSERT with duplicate order_id values causes error -> Option A

4. Quick Check:

   Duplicate primary keys cause insert errors [OK]

Hint: Duplicate primary keys cause insert errors [OK]

Common Mistakes:

• Assuming GENERATED BY DEFAULT forbids manual IDs
• Ignoring primary key uniqueness
• Thinking DEFAULT keyword causes errors

5.

You want to create a table employees with a IDENTITY column named id that automatically generates unique IDs but also allows you to insert specific IDs manually when needed. Which is the best way to define the ID column?

hard

A. Use SERIAL PRIMARY KEY for the ID column

B. Use INTEGER GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY

C. Use INTEGER GENERATED ALWAYS AS IDENTITY PRIMARY KEY

D. Use INTEGER without any auto-generation

Solution

1. Step 1: Understand difference between GENERATED ALWAYS and BY DEFAULT

   GENERATED ALWAYS AS IDENTITY disallows manual inserts; BY DEFAULT allows manual ID insertion.

2. Step 2: Match requirement to syntax

   Since manual insertion is needed sometimes, GENERATED BY DEFAULT AS IDENTITY is the best choice.

3. Final Answer:

   INTEGER GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY -> Option B

4. Quick Check:

   BY DEFAULT allows manual IDs, ALWAYS does not [OK]

Hint: BY DEFAULT allows manual ID insertion [OK]

Common Mistakes:

• Choosing GENERATED ALWAYS which forbids manual IDs
• Using SERIAL which is less flexible
• Not using any auto-generation and risking duplicates