Covering indexes with INCLUDE in PostgreSQL

Introduction

Covering indexes with INCLUDE help speed up queries by storing extra columns in the index, so the database can answer queries without looking at the main table.

When you want to make SELECT queries faster by avoiding extra table lookups.

When you have queries that filter by some columns but also select other columns.

When you want to reduce the time it takes to get results from large tables.

When you want to improve performance without adding too much extra storage.

When you want to optimize read-heavy workloads with specific query patterns.

Syntax

PostgreSQL

CREATE INDEX index_name ON table_name (column1, column2) INCLUDE (column3, column4);

The columns inside INCLUDE are not used for searching but are stored in the index for faster access.

This feature is specific to PostgreSQL and helps avoid extra table lookups.

Examples

Creates an index on last_name and stores first_name and email in the index for faster SELECT queries.

PostgreSQL

CREATE INDEX idx_users_name ON users (last_name) INCLUDE (first_name, email);

Indexes order_date and includes customer_id and total_amount to speed up queries selecting those columns.

PostgreSQL

CREATE INDEX idx_orders_date ON orders (order_date) INCLUDE (customer_id, total_amount);

Sample Program

This example creates a table employees, inserts some data, then creates a covering index on last_name including first_name and department. The SELECT query can use the index to get all needed columns without reading the main table.

PostgreSQL

CREATE TABLE employees (
  id SERIAL PRIMARY KEY,
  last_name TEXT NOT NULL,
  first_name TEXT NOT NULL,
  department TEXT NOT NULL,
  salary INT NOT NULL
);

INSERT INTO employees (last_name, first_name, department, salary) VALUES
('Smith', 'John', 'Sales', 50000),
('Doe', 'Jane', 'Marketing', 60000),
('Brown', 'Charlie', 'Sales', 55000);

CREATE INDEX idx_employees_lastname ON employees (last_name) INCLUDE (first_name, department);

-- Query that benefits from the covering index
SELECT last_name, first_name, department FROM employees WHERE last_name = 'Smith';

Important Notes

Including columns in the index does not affect the order of the index, only the columns used for searching do.

Covering indexes can make indexes larger, so use INCLUDE only for columns needed in SELECT but not in WHERE or JOIN.

PostgreSQL uses covering indexes automatically when possible to speed up queries.

Summary

Covering indexes store extra columns in the index to speed up SELECT queries.

Use INCLUDE to add columns that are selected but not searched on.

They reduce the need to read the main table, improving performance.

Practice

1. What is the main purpose of using INCLUDE in a PostgreSQL index?

easy

A. To change the data type of indexed columns

B. To create a unique constraint on the indexed columns

C. To delete columns from the index

D. To add extra columns to the index for faster SELECT queries without searching on them

Solution

1. Step 1: Understand the role of INCLUDE in indexes

   INCLUDE adds extra columns to the index that are not used for searching but can be returned in queries.

2. Step 2: Identify the benefit of these extra columns

   These extra columns help avoid reading the main table, speeding up SELECT queries that need those columns.

3. Final Answer:

   To add extra columns to the index for faster SELECT queries without searching on them -> Option D

4. Quick Check:

   INCLUDE adds columns for SELECT speed [OK]

Hint: INCLUDE adds columns to speed SELECT, not for searching [OK]

Common Mistakes:

• Thinking INCLUDE creates unique constraints
• Believing INCLUDE removes columns
• Assuming INCLUDE changes data types

2. Which of the following is the correct syntax to create a covering index on table users for column email and include last_login?

easy

A. CREATE INDEX idx_email ON users(email) INCLUDE (last_login);

B. CREATE INDEX idx_email ON users(email, last_login);

C. CREATE INDEX idx_email ON users INCLUDE (email, last_login);

D. CREATE INDEX idx_email ON users(email) WITH (last_login);

Solution

1. Step 1: Recall the syntax for INCLUDE in PostgreSQL indexes

   The correct syntax is to specify indexed columns first, then use INCLUDE for extra columns.

2. Step 2: Match the syntax to the options

   CREATE INDEX idx_email ON users(email) INCLUDE (last_login); correctly uses CREATE INDEX idx_email ON users(email) INCLUDE (last_login);

3. Final Answer:

   CREATE INDEX idx_email ON users(email) INCLUDE (last_login); -> Option A

4. Quick Check:

   Indexed columns first, INCLUDE for extras [OK]

Hint: Indexed columns before INCLUDE clause [OK]

Common Mistakes:

• Putting all columns inside parentheses without INCLUDE
• Using WITH instead of INCLUDE
• Including columns in wrong order

3. Given the index CREATE INDEX idx_name ON employees(last_name) INCLUDE (first_name, department);, what will the query SELECT last_name, first_name FROM employees WHERE last_name = 'Smith'; most likely do?

medium

A. Use the index but still access the table to get first_name

B. Use the index to find rows and return both last_name and first_name without accessing the table

C. Scan the whole table because first_name is not indexed

D. Return an error because first_name is not indexed

Solution

1. Step 1: Understand what INCLUDE columns do in the index

   INCLUDE columns are stored in the index to avoid accessing the main table for those columns.

2. Step 2: Analyze the query and index usage

   The query filters on last_name (indexed) and selects first_name (included). The index covers both, so no table access needed.

3. Final Answer:

   Use the index to find rows and return both last_name and first_name without accessing the table -> Option B

4. Quick Check:

   INCLUDE columns avoid table access [OK]

Hint: INCLUDE columns can be returned without table access [OK]

Common Mistakes:

• Assuming INCLUDE columns are not stored in the index
• Thinking table scan is always needed
• Confusing INCLUDE with indexed columns

4. You wrote this index: CREATE INDEX idx_order ON orders(order_date) INCLUDE (customer_id; but get a syntax error. What is the problem?

medium

A. INCLUDE cannot be used with order_date

B. You must list customer_id before order_date

C. Missing closing parenthesis after customer_id

D. INCLUDE requires at least two columns

Solution

1. Step 1: Check the syntax of the CREATE INDEX statement

   The statement has an opening parenthesis after INCLUDE but no closing parenthesis.

2. Step 2: Identify the syntax error

   Missing closing parenthesis causes the syntax error.

3. Final Answer:

   Missing closing parenthesis after customer_id -> Option C

4. Quick Check:

   Parentheses must be balanced [OK]

Hint: Check parentheses carefully in INCLUDE clause [OK]

Common Mistakes:

• Forgetting closing parenthesis
• Misordering columns
• Thinking INCLUDE needs multiple columns

5. You want to speed up this query: SELECT product_id, price, stock FROM products WHERE product_id = 123; by creating a covering index. Which index is best?

hard

A. CREATE INDEX idx_product ON products(product_id) INCLUDE (price, stock);

B. CREATE INDEX idx_product ON products(price, stock) INCLUDE (product_id);

C. CREATE INDEX idx_product ON products(product_id, price, stock);

D. CREATE INDEX idx_product ON products(product_id);

Solution

1. Step 1: Identify the filtering and selected columns in the query

   The query filters on product_id and selects price and stock.

2. Step 2: Choose an index that filters on product_id and includes price and stock

   CREATE INDEX idx_product ON products(product_id) INCLUDE (price, stock); indexes product_id and includes price and stock, covering the query efficiently.

3. Step 3: Compare other options

   CREATE INDEX idx_product ON products(price, stock) INCLUDE (product_id); indexes price and stock, not filtering column; CREATE INDEX idx_product ON products(product_id, price, stock); indexes all columns but includes unnecessary columns in index key; CREATE INDEX idx_product ON products(product_id); lacks included columns, so table access needed.

4. Final Answer:

   CREATE INDEX idx_product ON products(product_id) INCLUDE (price, stock); -> Option A

5. Quick Check:

   Filter column indexed, others included [OK]

Hint: Index filter column, INCLUDE others for covering [OK]

Common Mistakes:

• Including filter columns instead of indexing them
• Indexing all columns as keys unnecessarily
• Not including selected columns causing table access