PostgreSQL - ROW_NUMBER Function

The PostgreSQL ROW_NUMBER function is a crucial part of window functions, enabling users to assign unique sequential integers to rows within a dataset. This function is invaluable for tasks such as ranking, pagination and identifying duplicates.

The ROW_NUMBER() function generates a unique number for each row in a result set, starting from 1 for the first row. This numbering is reset whenever the partition changes (if used with the PARTITION BY clause).

Features of ROW_NUMBER()

• Sequential Numbering: Each row in the result set is assigned a unique number based on the specified order.
  
• Partitioning: The function can partition the result set into subsets, allowing for separate numbering within each partition.
  
• Sorting: The order of the assigned numbers can be defined using the ORDER BY clause.

Syntax:

The syntax for the ROW_NUMBER() function is as follows:

ROW_NUMBER() OVER ( [PARTITION BY partition_expression] ORDER BY order_expression )

Explanation of Syntax:

• PARTITION BY partition_expression: This clause is optional and divides the result set into partitions to which the ROW_NUMBER() function is applied. Each partition is numbered independently.
  
• ORDER BY order_expression: This clause is mandatory and specifies the order in which rows are numbered within each partition.

When to Use ROW_NUMBER()

The ROW_NUMBER() function is particularly useful in the following scenarios:

• Pagination: When displaying large datasets, ROW_NUMBER() can help in retrieving a specific range of records.
  
• Ranking: It can be used to assign ranks to rows based on specific criteria.
  
• Removing Duplicates: It can help identify duplicates by numbering rows within groups.

Examples of PostgreSQL ROW_NUMBER()

Let's go through some practical examples to illustrate the use of the ROW_NUMBER() function.

Example 1: Basic Usage of ROW_NUMBER()

Consider a table named employees with the following data:

Query:

SELECT 
    id, 
    name, 
    department, 
    salary,
    ROW_NUMBER() OVER (ORDER BY salary DESC) AS row_num
FROM employees;

Output:

Explain this Example:
In this example, the ROW_NUMBER() function assigns a unique number to each employee based on their salary in descending order.

Example 2: Using ROW_NUMBER() with PARTITION BY Clause

Now, let's say we want to assign row numbers within each department.

Query:

SELECT 
    id, 
    name, 
    department, 
    salary,
    ROW_NUMBER() OVER (PARTITION BY department ORDER BY salary DESC) AS row_num
FROM employees;

Output:

Explain this Example:

In this example, the ROW_NUMBER() function partitions the employees by department and assigns a row number based on their salary within each department.

Example 3: Combining ROW_NUMBER() with Common Table Expressions (CTE)

Common Table Expressions (CTEs) can enhance the readability of complex queries. Let's say we want to find the highest-paid employee in each department.

Query:

WITH RankedEmployees AS (
    SELECT 
        id, 
        name, 
        department, 
        salary,
        ROW_NUMBER() OVER (PARTITION BY department ORDER BY salary DESC) AS row_num
    FROM employees
)
SELECT * 
FROM RankedEmployees 
WHERE row_num = 1;

Output:

Explain this Example:

In this example, the CTE RankedEmployees assigns row numbers to each employee within their respective departments, and the main query selects only those with the highest salary.

Features Of PostgreSQL ROW_NUMBER() Function

• ROW_NUMBER() is a window function, meaning it requires an OVER clause to define the window of rows it operates on.
  
• The ROW_NUMBER() function assigns a unique integer to each row starting from 1 without skipping any numbers, even if rows have the same values in the columns used for ordering.
  
• The PARTITION BY clause divides the result set into partitions to which the ROW_NUMBER() function is applied independently. Without PARTITION BY, the function treats the entire result set as a single partition.
  
• ROW_NUMBER() can be combined with other window functions such as RANK(), DENSE_RANK(), and NTILE() to achieve different types of ranking and partitioning.