Named windows with WINDOW clause in PostgreSQL - Time & Space Complexity

Time Complexity: Named windows with WINDOW clause

O(n log n)

Understanding Time Complexity

We want to understand how the time to run a query changes when using named windows with the WINDOW clause in PostgreSQL.

Specifically, how does naming windows affect the work the database does as data grows?

Scenario Under Consideration

Analyze the time complexity of the following query using named windows.

SELECT employee_id, department_id, salary,
       AVG(salary) OVER dept_win AS avg_dept_salary,
       RANK() OVER dept_win AS rank_in_dept
FROM employees
WINDOW dept_win AS (PARTITION BY department_id ORDER BY salary DESC);
    

This query calculates the average salary and rank of employees within each department using a named window called dept_win.

Identify Repeating Operations

Look for repeated work inside the query.

• Primary operation: The database processes each department's employees separately (partition) and orders them by salary.
• How many times: Once per department, for all employees in that department.

How Execution Grows With Input

The work depends on how many employees there are and how they are grouped by department.

Input Size (n)	Approx. Operations
10 employees	About 10 operations, grouped by departments
100 employees	About 100 operations, grouped by departments
1000 employees	About 1000 operations, grouped by departments

Pattern observation: The total work grows roughly in direct proportion to the number of employees.

Final Time Complexity

Time Complexity: O(n log n)

This means the time grows a bit faster than the number of employees because sorting is needed within each department.

Common Mistake

[X] Wrong: "Using a named window makes the query run faster because it avoids repeated work."

[OK] Correct: Naming a window just helps write the query cleaner; the database still does the sorting and partitioning once per window, so the main work stays the same.

Interview Connect

Understanding how window functions work with partitions and ordering helps you explain query performance clearly and shows you know how databases handle grouped calculations efficiently.

Self-Check

"What if we removed the ORDER BY clause inside the WINDOW definition? How would the time complexity change?"

Practice

1. What is the main purpose of using the WINDOW clause in PostgreSQL?

easy

A. To join multiple tables using window functions

B. To create temporary tables for complex queries

C. To define named window specifications for reuse in the query

D. To specify the order of rows in a SELECT statement

Solution

1. Step 1: Understand the role of WINDOW clause

   The WINDOW clause allows defining named window specifications once for reuse.

2. Step 2: Compare options with WINDOW clause purpose

   Options A, B, and C describe other SQL features, not the WINDOW clause.

3. Final Answer:

   To define named window specifications for reuse in the query -> Option C

4. Quick Check:

   WINDOW clause = named window reuse [OK]

Hint: WINDOW clause defines reusable window specs [OK]

Common Mistakes:

• Confusing WINDOW with temporary tables
• Thinking WINDOW orders rows directly
• Mixing WINDOW with JOIN operations

2. Which of the following is the correct syntax to define a named window called w1 that orders rows by date ascending?

easy

A. WINDOW w1 (ORDER BY date ASC)

B. WINDOW w1 AS (ORDER BY date ASC)

C. WINDOW w1 BY (ORDER date ASC)

D. WINDOW AS w1 ORDER BY date ASC

Solution

1. Step 1: Recall WINDOW clause syntax

   The correct syntax is WINDOW name AS (window_definition), e.g., WINDOW w1 AS (ORDER BY date ASC).

2. Step 2: Check each option's syntax

   WINDOW w1 AS (ORDER BY date ASC) matches correct syntax. Options A, B, and D have incorrect keyword order or missing AS.

3. Final Answer:

   WINDOW w1 AS (ORDER BY date ASC) -> Option B

4. Quick Check:

   WINDOW name AS (definition) syntax [OK]

Hint: Use WINDOW name AS (definition) syntax [OK]

Common Mistakes:

• Omitting AS keyword
• Using BY instead of ORDER BY
• Misplacing parentheses

3. Given the query:

SELECT department, employee, salary, AVG(salary) OVER w AS avg_salary
FROM employees
WINDOW w AS (PARTITION BY department ORDER BY employee)
ORDER BY department, employee;

What does the avg_salary column show?

medium

A. The average salary of employees in the same department up to the current employee

B. The average salary of all employees in the company

C. The average salary of employees partitioned by department without order

D. The average salary of employees ordered by employee name across all departments

Solution

1. Step 1: Analyze the WINDOW clause

   The window named w partitions rows by department and orders by employee.

2. Step 2: Understand AVG() OVER w behavior

   AVG(salary) OVER w calculates the average salary for rows in the same department up to the current employee in order.

3. Final Answer:

   The average salary of employees in the same department up to the current employee -> Option A

4. Quick Check:

   Partition + order = running average per department [OK]

Hint: Partition + order means running calculation per group [OK]

Common Mistakes:

• Ignoring ORDER BY in window
• Assuming average over entire table
• Confusing partition with no partition

4. Consider this query:

SELECT name, salary, SUM(salary) OVER w
FROM employees
WINDOW w AS (PARTITION BY department ORDER BY name BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW);

It returns an error. What is the likely cause?

medium

A. The window frame clause requires ROWS or RANGE keyword explicitly

B. The window frame clause is invalid without RANGE keyword

C. The WINDOW clause must come after ORDER BY clause

D. The window name cannot be used in the OVER clause

Solution

1. Step 1: Check window frame syntax

   The frame clause must specify ROWS or RANGE explicitly before frame boundaries.

2. Step 2: Identify error cause

   The query has ORDER BY name BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW, missing the required ROWS or RANGE keyword before BETWEEN.

3. Final Answer:

   The window frame clause requires ROWS or RANGE keyword explicitly -> Option A

4. Quick Check:

   Window frame needs ROWS or RANGE keyword [OK]

Hint: Always specify ROWS or RANGE in window frame [OK]

Common Mistakes:

• Placing WINDOW clause after ORDER BY
• Omitting ROWS or RANGE keyword
• Using window name incorrectly in OVER

5. You want to calculate the cumulative sum of sales per region ordered by date, and reuse this window in multiple queries. Which is the best way to define and use a named window for this purpose?

hard

A. Define WINDOW sales_win AS (PARTITION BY region ORDER BY date ROWS UNBOUNDED PRECEDING) and use SUM(sales) OVER sales_win

B. Use SUM(sales) OVER (PARTITION BY region ORDER BY date) without WINDOW clause

C. Define WINDOW sales_win AS (ORDER BY date PARTITION BY region) and use SUM(sales) OVER sales_win

D. Define WINDOW sales_win AS (PARTITION BY region ORDER BY date RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) and use SUM(sales) OVER sales_win

Solution

1. Step 1: Understand cumulative sum window frame

   Cumulative sum requires a frame from the start to current row, so use RANGE or ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW.

2. Step 2: Check window definition correctness

   Define WINDOW sales_win AS (PARTITION BY region ORDER BY date RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) and use SUM(sales) OVER sales_win correctly defines WINDOW with PARTITION BY region, ORDER BY date, and RANGE frame for cumulative sum.

3. Step 3: Compare with other options

   Define WINDOW sales_win AS (PARTITION BY region ORDER BY date ROWS UNBOUNDED PRECEDING) and use SUM(sales) OVER sales_win uses ROWS UNBOUNDED PRECEDING which sums the entire partition, not cumulatively. Define WINDOW sales_win AS (ORDER BY date PARTITION BY region) and use SUM(sales) OVER sales_win has wrong order of clauses. Use SUM(sales) OVER (PARTITION BY region ORDER BY date) without WINDOW clause does not reuse named window.

4. Final Answer:

   Define WINDOW sales_win AS (PARTITION BY region ORDER BY date RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) and use SUM(sales) OVER sales_win -> Option D

5. Quick Check:

   Named window with correct frame = Define WINDOW sales_win AS (PARTITION BY region ORDER BY date RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) and use SUM(sales) OVER sales_win [OK]

Hint: Use RANGE UNBOUNDED PRECEDING for cumulative sums [OK]

Common Mistakes:

• Omitting frame clause for cumulative sum
• Incorrect order of PARTITION BY and ORDER BY
• Not using named window for reuse