NTILE for distribution in PostgreSQL

Introduction

NTILE helps split data into equal parts or groups. It shows which group each row belongs to.

You want to divide sales data into 4 equal groups to see top sellers.

You need to split students' scores into 3 groups: low, medium, high.

You want to assign customers into 5 groups for targeted marketing.

You want to rank employees into quartiles based on performance.

Syntax

PostgreSQL

NTILE(number_of_groups) OVER (ORDER BY column_name)

number_of_groups is how many parts you want to split the data into.

The ORDER BY inside OVER() decides how rows are sorted before splitting.

Examples

This splits students into 4 groups based on their scores, highest first.

PostgreSQL

SELECT name, score, NTILE(4) OVER (ORDER BY score DESC) AS quartile FROM students;

This divides products into 3 groups by sales, lowest to highest.

PostgreSQL

SELECT product, sales, NTILE(3) OVER (ORDER BY sales) AS tertile FROM sales_data;

Sample Program

This example creates a sales table, inserts 5 products with amounts, then splits them into 2 groups (halves) by amount.

PostgreSQL

CREATE TABLE sales (
  id SERIAL PRIMARY KEY,
  product VARCHAR(20),
  amount INT
);

INSERT INTO sales (product, amount) VALUES
('A', 100),
('B', 200),
('C', 300),
('D', 400),
('E', 500);

SELECT product, amount, NTILE(2) OVER (ORDER BY amount) AS half
FROM sales
ORDER BY amount;

Important Notes

If the number of rows can't be divided evenly, some groups will have one more row.

NTILE always assigns group numbers starting from 1 up to the number of groups.

Summary

NTILE splits rows into equal groups based on order.

Use it to create quartiles, tertiles, or any group count.

Groups are numbered starting at 1.

Practice

1. What does the NTILE(n) function do in PostgreSQL?

easy

A. It returns the maximum value of a column.

B. It counts the total number of rows in a table.

C. It divides rows into n groups with roughly equal size, numbering each group.

D. It sorts rows alphabetically by a specified column.

Solution

1. Step 1: Understand NTILE purpose

   NTILE(n) splits rows into n groups based on order.

2. Step 2: Recognize group numbering

   Each group is numbered starting from 1 to n, distributing rows evenly.

3. Final Answer:

   It divides rows into n groups with roughly equal size, numbering each group. -> Option C

4. Quick Check:

   NTILE splits rows into groups [OK]

Hint: NTILE splits rows into equal groups numbered 1 to n [OK]

Common Mistakes:

• Confusing NTILE with COUNT or MAX functions
• Thinking NTILE sorts rows alphabetically
• Assuming NTILE returns row counts

2. Which of the following is the correct syntax to assign quartiles using NTILE in PostgreSQL?

easy

A. SELECT column, NTILE(4) PARTITION BY column FROM table;

B. SELECT column, NTILE OVER (4 ORDER BY column) AS quartile FROM table;

C. SELECT column, NTILE(4) AS quartile FROM table ORDER BY column;

D. SELECT column, NTILE(4) OVER (ORDER BY column) AS quartile FROM table;

Solution

1. Step 1: Recall NTILE syntax

   The correct syntax is NTILE(n) OVER (ORDER BY column) to assign groups.

2. Step 2: Check each option

   SELECT column, NTILE(4) OVER (ORDER BY column) AS quartile FROM table; uses correct syntax with NTILE(4) OVER (ORDER BY column). Others misuse OVER clause or omit it.

3. Final Answer:

   SELECT column, NTILE(4) OVER (ORDER BY column) AS quartile FROM table; -> Option D

4. Quick Check:

   NTILE needs OVER with ORDER BY [OK]

Hint: NTILE requires OVER and ORDER BY clauses [OK]

Common Mistakes:

• Omitting OVER clause
• Placing number inside OVER incorrectly
• Using PARTITION BY without ORDER BY

3. Given the table scores with column score values: 10, 20, 30, 40, 50, what is the output of this query?

SELECT score, NTILE(3) OVER (ORDER BY score) AS tile FROM scores ORDER BY score;

medium

A. [{10,1}, {20,1}, {30,2}, {40,2}, {50,3}]

B. [{10,1}, {20,2}, {30,2}, {40,3}, {50,3}]

C. [{10,1}, {20,1}, {30,1}, {40,2}, {50,3}]

D. [{10,1}, {20,2}, {30,3}, {40,3}, {50,3}]

Solution

1. Step 1: Understand NTILE(3) distribution

   NTILE(3) divides 5 rows into 3 groups as evenly as possible: 2 rows in first group, 2 in second, 1 in third.

2. Step 2: Assign tiles by order

   Ordered scores: 10,20 (tile 1), 30,40 (tile 2), 50 (tile 3).

3. Final Answer:

   [{10,1}, {20,1}, {30,2}, {40,2}, {50,3}] -> Option A

4. Quick Check:

   5 rows split into 3 groups: 2,2,1 [OK]

Hint: Divide rows evenly; extra rows go to earlier tiles [OK]

Common Mistakes:

• Assigning one row per tile ignoring total rows
• Not ordering rows before applying NTILE
• Assuming equal rows per tile always

4. Identify the error in this query:

SELECT id, NTILE(4) ORDER BY score FROM results;

medium

A. Missing OVER() clause after NTILE(4).

B. ORDER BY cannot be used inside NTILE function.

C. NTILE requires a PARTITION BY clause.

D. The query is correct as is.

Solution

1. Step 1: Check NTILE syntax

   NTILE requires an OVER() clause with ORDER BY inside it, like NTILE(4) OVER (ORDER BY score).

2. Step 2: Identify missing OVER()

   The query misses the OVER() clause, causing syntax error.

3. Final Answer:

   Missing OVER() clause after NTILE(4). -> Option A

4. Quick Check:

   NTILE needs OVER() with ORDER BY [OK]

Hint: NTILE always needs OVER() with ORDER BY [OK]

Common Mistakes:

• Placing ORDER BY outside OVER()
• Forgetting OVER() clause
• Assuming PARTITION BY is mandatory

5. You want to divide 7 employees into 3 salary groups using NTILE, but the salaries are:
5000, 7000, 7000, 9000, 11000, 11000, 13000.
Which NTILE(3) group assignment is correct when ordered by salary?

hard

A. [{5000,1}, {7000,1}, {7000,1}, {9000,2}, {11000,2}, {11000,3}, {13000,3}]

B. [{5000,1}, {7000,1}, {7000,2}, {9000,3}, {11000,3}, {11000,3}, {13000,3}]

C. [{5000,1}, {7000,2}, {7000,2}, {9000,2}, {11000,3}, {11000,3}, {13000,3}]

D. [{5000,1}, {7000,1}, {7000,2}, {9000,2}, {11000,3}, {11000,3}, {13000,3}]

Solution

1. Step 1: Calculate group sizes for 7 rows and 3 groups

   7 rows divided into 3 groups: first group gets 3 rows, next two get 2 rows each.

2. Step 2: Assign groups by ordered salary

   Order: 5000(1), 7000(1), 7000(2), 9000(3), 11000(3), 11000(3), 13000(3).

3. Final Answer:

   [{5000,1}, {7000,1}, {7000,2}, {9000,3}, {11000,3}, {11000,3}, {13000,3}] -> Option B

4. Quick Check:

   7 rows split into 3 groups: 3,2,2 [OK]

Hint: Extra rows go to earlier groups; order matters [OK]

Common Mistakes:

• Ignoring duplicate salaries affecting group boundaries
• Assigning equal rows per group ignoring total rows
• Misplacing group numbers for duplicates