Why window functions are needed in SQL - See It in Action

Understanding Why Window Functions Are Needed in SQL

📖 Scenario: You work as a data analyst for a retail company. You have a sales table with daily sales data for different stores. You want to analyze sales trends and compare each day's sales to the average sales of the store without losing the detail of each row.

🎯 Goal: Build a simple SQL query using window functions to calculate the average sales per store alongside each day's sales.

📋 What You'll Learn

Create a table called sales with columns store_id, sale_date, and amount.

Insert sample data for at least two stores with sales on multiple dates.

Write a query that uses a window function to calculate the average sales per store.

Show each sale amount along with the average sales for that store.

💡 Why This Matters

🌍 Real World

Window functions are used in business reports to compare individual records with group statistics, like sales per day compared to average sales per store.

💼 Career

Data analysts and database developers use window functions to write efficient queries that provide detailed and summarized data together, improving insights without complex joins or subqueries.

1

Create the sales table and insert data

Create a table called sales with columns store_id (integer), sale_date (date), and amount (integer). Insert these exact rows: (1, '2024-01-01', 100), (1, '2024-01-02', 150), (2, '2024-01-01', 200), (2, '2024-01-02', 250).

SQL

-- Create the sales table and insert the specified rows
-- Your code here

Hint

Use CREATE TABLE to define the table and INSERT INTO to add rows.

2

Write a basic SELECT query

Write a SELECT query to retrieve all columns from the sales table ordered by store_id and sale_date.

SQL

CREATE TABLE sales (
  store_id INTEGER,
  sale_date DATE,
  amount INTEGER
);

INSERT INTO sales (store_id, sale_date, amount) VALUES
(1, '2024-01-01', 100),
(1, '2024-01-02', 150),
(2, '2024-01-01', 200),
(2, '2024-01-02', 250);

-- Write a SELECT query to get all columns ordered by store_id and sale_date
-- Your code here

Hint

Use ORDER BY store_id, sale_date to sort the results.

3

Add a window function to calculate average sales per store

Modify the SELECT query to add a column called avg_sales that shows the average amount per store_id using the window function AVG(amount) OVER (PARTITION BY store_id).

SQL

CREATE TABLE sales (
  store_id INTEGER,
  sale_date DATE,
  amount INTEGER
);

INSERT INTO sales (store_id, sale_date, amount) VALUES
(1, '2024-01-01', 100),
(1, '2024-01-02', 150),
(2, '2024-01-01', 200),
(2, '2024-01-02', 250);

-- Add a column avg_sales using AVG(amount) OVER (PARTITION BY store_id)
-- Your code here

Hint

Use AVG(amount) OVER (PARTITION BY store_id) to calculate the average per store.

4

Explain why window functions are useful here

Add a comment explaining that window functions let you calculate aggregates like averages without losing the detail of each row, unlike GROUP BY which groups rows together.

SQL

CREATE TABLE sales (
  store_id INTEGER,
  sale_date DATE,
  amount INTEGER
);

INSERT INTO sales (store_id, sale_date, amount) VALUES
(1, '2024-01-01', 100),
(1, '2024-01-02', 150),
(2, '2024-01-01', 200),
(2, '2024-01-02', 250);

SELECT store_id, sale_date, amount,
       AVG(amount) OVER (PARTITION BY store_id) AS avg_sales
FROM sales
ORDER BY store_id, sale_date;

-- Add a comment explaining why window functions are needed
-- Your code here

Hint

Explain that window functions keep row details while adding aggregate info.

Practice

1. Why are window functions needed in SQL?
SELECT with window functions can do what that regular aggregation cannot?

easy

A. They speed up query execution by skipping indexes.

B. They allow calculations across rows without hiding individual rows.

C. They automatically create new tables for results.

D. They replace the need for JOIN operations.

Solution

1. Step 1: Understand aggregation vs window functions

   Regular aggregation groups rows and returns fewer rows, hiding details.

2. Step 2: Identify window function behavior

   Window functions calculate values across rows but keep all original rows visible.

3. Final Answer:

   They allow calculations across rows without hiding individual rows. -> Option B

4. Quick Check:

   Window functions keep rows visible = A [OK]

Hint: Window functions keep all rows, aggregations reduce rows [OK]

Common Mistakes:

• Thinking window functions create new tables
• Confusing window functions with JOINs
• Believing window functions speed up queries

2. Which of the following is the correct syntax to use a window function in SQL?

easy

A. SELECT name, RANK() OVER (ORDER BY score) FROM players;

B. SELECT name, RANK(score) FROM players ORDER BY score;

C. SELECT name, RANK() FROM players GROUP BY score;

D. SELECT name, RANK OVER (ORDER BY score) FROM players;

Solution

1. Step 1: Recall window function syntax

   Window functions use function() OVER (PARTITION BY ... ORDER BY ...).

2. Step 2: Check each option

   SELECT name, RANK() OVER (ORDER BY score) FROM players; correctly uses RANK() OVER (ORDER BY score). Others miss parentheses or OVER keyword.

3. Final Answer:

   SELECT name, RANK() OVER (ORDER BY score) FROM players; -> Option A

4. Quick Check:

   Window function syntax = function() OVER (...) [OK]

Hint: Window functions always use OVER() clause [OK]

Common Mistakes:

• Omitting parentheses after function name
• Forgetting the OVER keyword
• Using GROUP BY instead of window functions

3. Given the table sales with columns id, region, and amount, what will this query return?

SELECT id, region, amount, SUM(amount) OVER (PARTITION BY region ORDER BY id) AS running_total FROM sales ORDER BY id;

medium

A. A running total of sales amount per region, ordered by id, showing all rows.

B. Total sales amount per region, one row per region only.

C. Sum of all sales amounts ignoring region, one total value.

D. Syntax error due to missing GROUP BY clause.

Solution

1. Step 1: Understand SUM() OVER with PARTITION and ORDER

   SUM() OVER (PARTITION BY region ORDER BY id) calculates a running total per region ordered by id.

2. Step 2: Check output rows

   Window functions keep all rows, so all sales rows appear with running totals.

3. Final Answer:

   A running total of sales amount per region, ordered by id, showing all rows. -> Option A

4. Quick Check:

   Running total per group = A [OK]

Hint: SUM() OVER with PARTITION BY gives running totals per group [OK]

Common Mistakes:

• Thinking it returns one row per group
• Confusing window function with GROUP BY aggregation
• Expecting syntax error without GROUP BY

4. Identify the error in this SQL query using a window function:

SELECT employee_id, department, ROW_NUMBER() ORDER BY salary FROM employees;

medium

A. ORDER BY cannot be used inside ROW_NUMBER().

B. employee_id must be inside ROW_NUMBER() function.

C. ROW_NUMBER() requires GROUP BY clause.

D. Missing OVER() clause after ROW_NUMBER() function.

Solution

1. Step 1: Check window function syntax

   ROW_NUMBER() must be followed by OVER() with optional ORDER BY inside it.

2. Step 2: Identify missing OVER()

   The query uses ROW_NUMBER() ORDER BY salary without OVER(), causing syntax error.

3. Final Answer:

   Missing OVER() clause after ROW_NUMBER() function. -> Option D

4. Quick Check:

   Window functions need OVER() [OK]

Hint: Always add OVER() after window functions [OK]

Common Mistakes:

• Placing ORDER BY outside OVER()
• Thinking GROUP BY is needed for ROW_NUMBER()
• Misunderstanding function arguments

5. You want to find each employee's salary rank within their department, but also show all employees with their salaries and ranks. Which SQL approach best achieves this?

hard

A. Use a subquery to calculate ranks and join back to employees table.

B. Use GROUP BY department, salary with RANK() to get ranks per department.

C. Use RANK() OVER (PARTITION BY department ORDER BY salary DESC) to rank salaries per department while keeping all rows.

D. Use ORDER BY salary DESC with LIMIT to get top salaries per department.

Solution

1. Step 1: Understand the requirement

   We want ranks per department and all employees listed with their ranks.

2. Step 2: Evaluate options

   Use RANK() OVER (PARTITION BY department ORDER BY salary DESC) to rank salaries per department while keeping all rows. uses RANK() OVER (PARTITION BY department ORDER BY salary DESC) which ranks salaries per department and keeps all rows visible.

3. Final Answer:

   Use RANK() OVER (PARTITION BY department ORDER BY salary DESC) to rank salaries per department while keeping all rows. -> Option C

4. Quick Check:

   Window function with PARTITION keeps all rows and ranks [OK]

Hint: Use RANK() OVER PARTITION BY to rank within groups [OK]

Common Mistakes:

• Using GROUP BY loses individual rows
• Using subqueries unnecessarily complicates query
• LIMIT filters rows, not ranks all