Composite types in PostgreSQL

Introduction

Composite types let you group different pieces of data together as one unit. This helps organize related information clearly.

When you want to store multiple related values in one column, like a full address with street, city, and zip.

When you want to return multiple related values from a function in a neat package.

When you want to pass a group of related values as one argument to a function.

When you want to create a reusable data structure inside the database.

When you want to simplify complex queries by grouping related fields.

Syntax

PostgreSQL

CREATE TYPE type_name AS (
    field1 data_type1,
    field2 data_type2
);

Composite types are like custom data structures made inside the database.

You can use them as column types or function return types.

Examples

This creates a composite type named full_name with two text fields.

PostgreSQL

CREATE TYPE full_name AS (
    first_name TEXT,
    last_name TEXT
);

This defines an address type with street, city, and zip code fields.

PostgreSQL

CREATE TYPE address AS (
    street TEXT,
    city TEXT,
    zip_code TEXT
);

This table uses the composite types full_name and address as column types.

PostgreSQL

CREATE TABLE people (
    id SERIAL PRIMARY KEY,
    name full_name,
    home address
);

Sample Program

This example creates a composite type full_name, then a table employees using it. It inserts two employees and selects their names by accessing the composite fields.

PostgreSQL

CREATE TYPE full_name AS (
    first_name TEXT,
    last_name TEXT
);

CREATE TABLE employees (
    id SERIAL PRIMARY KEY,
    name full_name
);

INSERT INTO employees (name) VALUES (ROW('Alice', 'Smith'));
INSERT INTO employees (name) VALUES (ROW('Bob', 'Jones'));

SELECT id, (name).first_name AS first_name, (name).last_name AS last_name FROM employees ORDER BY id;

Important Notes

You access composite fields using parentheses and dot notation, like (column).field.

Composite types can be nested inside other composite types.

They help keep related data together and make queries easier to read.

Summary

Composite types group related fields into one custom type.

They can be used as column types or function return types.

Access fields with (column).field syntax.

Practice

1. What is a composite type in PostgreSQL?

easy

A. A special index type for faster queries

B. A built-in type for storing large text data

C. A custom type that groups multiple related fields together

D. A function that returns multiple rows

Solution

1. Step 1: Understand the definition of composite types

   Composite types are user-defined types that group several fields into one unit, like a record.

2. Step 2: Compare with other options

   Options A, B, and D describe unrelated PostgreSQL features, not composite types.

3. Final Answer:

   A custom type that groups multiple related fields together -> Option C

4. Quick Check:

   Composite type = grouped fields [OK]

Hint: Composite types group fields like a mini table row [OK]

Common Mistakes:

• Confusing composite types with arrays
• Thinking composite types are indexes
• Mixing composite types with functions

2. Which of the following is the correct syntax to define a composite type named person with fields name (text) and age (integer)?

easy

A. CREATE TYPE person AS (name text, age integer);

B. CREATE TABLE person (name text, age integer);

C. CREATE TYPE person (name text, age integer);

D. CREATE COMPOSITE person AS (name text, age integer);

Solution

1. Step 1: Recall the syntax for creating composite types

   The correct syntax uses CREATE TYPE with AS and parentheses listing fields and types.

2. Step 2: Check each option

   CREATE TYPE person AS (name text, age integer); matches the correct syntax. CREATE TABLE person (name text, age integer); creates a table, not a type. CREATE TYPE person (name text, age integer); misses AS keyword. CREATE COMPOSITE person AS (name text, age integer); uses invalid keyword COMPOSITE.

3. Final Answer:

   CREATE TYPE person AS (name text, age integer); -> Option A

4. Quick Check:

   CREATE TYPE ... AS (...) is correct [OK]

Hint: Use CREATE TYPE name AS (fields) for composite types [OK]

Common Mistakes:

• Using CREATE TABLE instead of CREATE TYPE
• Omitting AS keyword
• Using non-existent COMPOSITE keyword

3. Given the composite type and table:
CREATE TYPE address AS (city text, zip integer);
CREATE TABLE users (id serial, home address);
What will the query return?
SELECT (home).city FROM users WHERE id = 1;

medium

A. An error because you cannot access composite fields like this

B. The zip code of the home address for user with id 1

C. The entire home composite value as text

D. The city name stored in the home column for user with id 1

Solution

1. Step 1: Understand composite field access syntax

   Using (home).city extracts the city field from the composite column home.

2. Step 2: Analyze the query result

   The query selects city from home for user with id 1, so it returns that city name.

3. Final Answer:

   The city name stored in the home column for user with id 1 -> Option D

4. Quick Check:

   (column).field extracts field from composite [OK]

Hint: Use (column).field to get composite field value [OK]

Common Mistakes:

• Trying to access composite fields without parentheses
• Expecting entire composite instead of single field
• Confusing city with zip field

4. Identify the error in this code snippet:
CREATE TYPE product_info AS (name text, price numeric);
CREATE TABLE products (id serial, info product_info);
INSERT INTO products (info) VALUES ('Laptop', 999.99);

medium

A. The serial type is invalid for id column

B. The INSERT statement syntax is incorrect for composite type

C. The table products should not use composite types

D. The CREATE TYPE statement is missing a semicolon

Solution

1. Step 1: Check INSERT syntax for composite types

   When inserting a composite type, values must be wrapped in parentheses or ROW(), not as separate values.

2. Step 2: Identify the error in the INSERT statement

   The statement tries to insert two separate values without wrapping them as a composite, causing syntax error.

3. Final Answer:

   The INSERT statement syntax is incorrect for composite type -> Option B

4. Quick Check:

   Insert composite as (value1, value2) or ROW(...) [OK]

Hint: Wrap composite values in parentheses or use ROW() [OK]

Common Mistakes:

• Inserting composite fields as separate values
• Forgetting parentheses around composite values
• Misunderstanding serial type usage

5. You have a composite type location with fields latitude and longitude. You want to create a function that returns this composite type and use it in a query. Which of the following is the correct way to define the function?

hard

A. CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN ROW(40.7128, -74.0060); END; $$ LANGUAGE plpgsql;

B. CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN 40.7128, -74.0060; END; $$ LANGUAGE plpgsql;

C. CREATE FUNCTION get_location() RETURNS TABLE(latitude float, longitude float) AS $$ BEGIN RETURN (40.7128, -74.0060); END; $$ LANGUAGE plpgsql;

D. CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN '40.7128, -74.0060'; END; $$ LANGUAGE plpgsql;

Solution

1. Step 1: Understand how to return composite types from functions

   Functions returning composite types must return a row value, typically using ROW() constructor.

2. Step 2: Analyze each option

   CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN ROW(40.7128, -74.0060); END; $$ LANGUAGE plpgsql; correctly uses RETURN ROW(...) to return the composite type. CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN 40.7128, -74.0060; END; $$ LANGUAGE plpgsql; returns multiple values without row constructor (parentheses or ROW()), causing syntax error. CREATE FUNCTION get_location() RETURNS TABLE(latitude float, longitude float) AS $$ BEGIN RETURN (40.7128, -74.0060); END; $$ LANGUAGE plpgsql; returns a table, not the composite type. CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN '40.7128, -74.0060'; END; $$ LANGUAGE plpgsql; returns a string, not composite.

3. Final Answer:

   CREATE FUNCTION get_location() RETURNS location AS $$ BEGIN RETURN ROW(40.7128, -74.0060); END; $$ LANGUAGE plpgsql; -> Option A

4. Quick Check:

   Use RETURN ROW(...) for composite return [OK]

Hint: Use RETURN ROW(...) to return composite types in functions [OK]

Common Mistakes:

• Returning values without row constructor
• Returning string instead of composite
• Confusing RETURNS TABLE with RETURNS composite