Window Functions in SQL - How they work?

with examples
What is a Window Function?
A Window Function is used to perform a calculation on an aggregate value based on a set of rows and return multiple rows for each group.

AGGREGATE FUNCTIONS WINDOW FUNCTIONS

Types of Window Fucntions

Examples Concept
Example-1 COUNT
Example-2 SUM
Window Example-3 AVG
Functions Example-4 ROW_NUMBER
Example-5 RANK
Example-6 DENSE_RANK
Example-7 NTILE
Aggregate Ranking Value Example-8 LEAD
Functions Functions Functions Example-9 LAG
Example-10FIRST_VALUE
Example-11LAST_VLAUE
COUNT() ROW_NUMBER() LAG()
SUM() RANK() LEAD()
AVG() DENSE_RANK() FIRST_VALUE()
CUME_DIST() LAST_VALUE()
NTILE()
1 Find the the total number of clients per state and city

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's first partition the data by states and count the number of rows in each partitio

state city client_id sales no_of_clients

California Los Angeles 1001 10000 4
California Los Angeles 1002 10000 4
California San Diego 1003 30000 4
California San Diego 1004 40000 4
Texas Houston 1005 60000 4
Texas Houston 1006 50000 4
Texas Austin 1007 60000 4
Texas Austin 1008 50000 4

state no_of_clients
California 4
California 4
California 4
California 4
Texas 4
Texas 4
Texas 4
Texas 4

Now, let's partition the data by cities and count the number of rows in each partitio

state city client_id sales no_of_clients

California Los Angeles 1001 10000 2
 California Los Angeles 1002 10000 2
California San Diego 1003 30000 2
California San Diego 1004 40000 2
Texas Houston 1005 60000 2
Texas Houston 1006 50000 2
Texas Austin 1007 60000 2
Texas Austin 1008 50000 2

city no_of_clients
Los Angeles 2
Los Angeles 2
San Diego 2
San Diego 2
Houston 2
Houston 2
Austin 2
Austin 2

How can I write a query to get this answer programmatically?

COUNT(client_id) OVER(Partition by city) as clients_city
FROM revenue;

client_id state city clients_stateclients_city

1001 California Los Angeles 4 2
1002 California Los Angeles 4 2
1003 California San Diego 4 2
1004 California San Diego 4 2
1005 Texas Houston 4 2
1006 Texas Houston 4 2
1007 Texas Austin 4 2
1008 Texas Austin 4 2
er of rows in each partition

er of rows in each partition

2 Write a query to show the total amount of sales in each state and the total amount

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's first partition the data by states and find the sum of sales in each partition

state city client_id sales state_sales

California Los Angeles 1001 10000 90000
California Los Angeles 1002 10000 90000
California San Diego 1003 30000 90000
California San Diego 1004 40000 90000
Texas Houston 1005 60000 220000
Texas Houston 1006 50000 220000
Texas Austin 1007 60000 220000
Texas Austin 1008 50000 220000

state state_sales
California 90000
California 90000
California 90000
California 90000
Texas 220000
Texas 220000
Texas 220000
Texas 220000

Now, let's partition the data by cities and find the sum of sales in each partition

state city client_id sales city_sales

California Los Angeles 1001 10000 20000
 California Los Angeles 1002 10000 20000
California San Diego 1003 30000 70000
California San Diego 1004 40000 70000
Texas Houston 1005 60000 110000
Texas Houston 1006 50000 110000
Texas Austin 1007 60000 110000
Texas Austin 1008 50000 110000

city city_sales
Los Angeles 20000
Los Angeles 20000
San Diego 70000
San Diego 70000
Houston 110000
Houston 110000
Austin 110000
Austin 110000

How can I write a query to get this answer programmatically?

FROM revenue
ORDER BY client_id;

state city state_sales city_sales

California Los Angeles 90000 20000
California Los Angeles 90000 20000
California San Diego 90000 70000
California San Diego 90000 70000
Texas Houston 220000 110000
Texas Houston 220000 110000
Texas Austin 220000 110000
Texas Austin 220000 110000
ate and the total amount of sales in each city

sales in each partition

sales in each partition

3 Write a query to show the average amount of sales in each state and the average a

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's first partition the data by states and find the average amount of sales in each

state city client_id sales avg_state

California Los Angeles 1001 10000 22500
California Los Angeles 1002 10000 22500
California San Diego 1003 30000 22500
California San Diego 1004 40000 22500
Texas Houston 1005 60000 55000
Texas Houston 1006 50000 55000
Texas Austin 1007 60000 55000
Texas Austin 1008 50000 55000

state avg_state
California 22500
California 22500
California 22500
California 22500
Texas 55000
Texas 55000
Texas 55000
Texas 55000

Now, let's partition the data by cities and find the average amount of sales in each

state city client_id sales city_sales

California Los Angeles 1001 10000 10000
 California Los Angeles 1002 10000 10000
California San Diego 1003 30000 35000
California San Diego 1004 40000 35000
Texas Houston 1005 60000 55000
Texas Houston 1006 50000 55000
Texas Austin 1007 60000 55000
Texas Austin 1008 50000 55000

city avg_city
Los Angeles 10000
Los Angeles 10000
San Diego 35000
San Diego 35000
Houston 55000
Houston 55000
Austin 55000
Austin 55000

How can I write a query to get this answer programmatically?

AVG(sales) OVER(Partition by city) as avg_city
FROM revenue;

client_id state city avg_state avg_city

1001 California Los Angeles 22500 10000
1002 California Los Angeles 22500 10000
1003 California San Diego 22500 35000
1004 California San Diego 22500 35000
1005 Texas Houston 55000 55000
1006 Texas Houston 55000 55000
1007 Texas Austin 55000 55000
1008 Texas Austin 55000 55000
h state and the average amount of sales in each city

e amount of sales in each partition

amount of sales in each partition

4 Write a query to assign row numbers to the revenue table based on states

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's first partition the data by states and assign row numbers to each row in the p

state city client_id sales row_no

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 2
California San Diego 1003 30000 3
California San Diego 1004 40000 4
Texas Houston 1005 60000 1
Texas Houston 1006 50000 2
Texas Austin 1007 60000 3
Texas Austin 1008 50000 4

state row_no
California 1
California 2
California 3
California 4
Texas 1
Texas 2
Texas 3
Texas 4

How can I write a query to get this answer programmatically?

 ROW_NUMBER() OVER(Partition by state) as row_no
FROM revenue;

state city client_id sales row_no

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 2
California San Diego 1003 30000 3
California San Diego 1004 40000 4
Texas Houston 1005 60000 1
Texas Houston 1006 50000 2
Texas Austin 1007 60000 3
Texas Austin 1008 50000 4
based on states

bers to each row in the partition

5 Write a query to assign rank based on the sales in each state

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's rank the data by partitioning the states and sorting sales in ascending order

state city client_id sales rnk

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 3
California San Diego 1004 40000 4
Texas Houston 1006 50000 1
Texas Austin 1008 50000 1
Texas Houston 1005 60000 3
Texas Austin 1007 60000 3

How can I write a query to get this answer programmatically?

RANK() OVER(Partition by state ORDER BY sales) as rnk
FROM revenue;

state city client_id sales rnk

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 3
California San Diego 1004 40000 4
Texas Houston 1006 50000 1
Texas Austin 1008 50000 1
Texas Houston 1005 60000 3
Texas Austin 1007 60000 3
sales in ascending order
6 Write a query to assign rank based on the sales of each state

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's rank the data by partitioning the states and sorting sales in ascending order

state city client_id sales dense_rnk

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 2
California San Diego 1004 40000 3
Texas Houston 1006 50000 1
Texas Austin 1008 50000 1
Texas Houston 1005 60000 2
Texas Austin 1007 60000 2

How can I write a query to get this answer programmatically?

DENSE_RANK() OVER(Partition by state ORDER BY sales) as dense_rnk
FROM revenue;

state city client_id sales dense_rnk

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 2
California San Diego 1004 40000 3
Texas Houston 1006 50000 1
Texas Austin 1008 50000 1
Texas Houston 1005 60000 2
Texas Austin 1007 60000 2
sales in ascending order

NOTE:
) as dense_rnk
Unlike the RANK() function, the
DENSE_RANK() function returns
consecutive rank values.
8 Write a query to bucket the data into 3 buckets based on the range of sales

Revenue
state city client_id sales
California Los Angeles 1001 10000
California Los Angeles 1002 10000
California San Diego 1003 30000
California San Diego 1004 40000
Texas Houston 1005 60000
Texas Houston 1006 50000
Texas Austin 1007 60000
Texas Austin 1008 50000

Let's first order the data by sales and then group them into three groups based on

state city client_id sales tier

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 1
California San Diego 1004 40000 2
Texas Houston 1006 50000 2
Texas Austin 1008 50000 2
Texas Houston 1005 60000 3
Texas Austin 1007 60000 3

How can I write a query to get this answer programmatically?

NTILE(3) OVER(ORDER BY sales) as tier
FROM revenue;

client_id state city sales tier

California Los Angeles 1001 10000 1
California Los Angeles 1002 10000 1
California San Diego 1003 30000 1
California San Diego 1004 40000 2
Texas Houston 1006 50000 2
Texas Austin 1008 50000 2
Texas Houston 1005 60000 3
Texas Austin 1007 60000 3
he range of sales

o three groups based on the range of sales

9 Write a query to find the sales of the following month in each of the states

Revenue
state month sales
California 4 30000
California 1 10000
California 3 40000
California 2 10000
Texas 1 50000
Texas 4 60000
Texas 3 60000
Texas 2 50000

Let's first partition the states and then sort the data by month

state month sales

California 1 10000
California 2 10000
California 3 40000
California 4 30000
Texas 1 50000
Texas 2 50000
Texas 3 60000
Texas 4 60000

Now, let's find the sales of the following month

state month sales next_sales

California 1 10000 10000
California 2 10000 40000
California 3 40000 30000
California 4 30000 Null
Texas 1 50000 50000
Texas 2 50000 60000
Texas 3 60000 60000
Texas 4 60000 Null

How can I write a query to get this answer programmatically?

 LEAD(sales) OVER(PARTITION BY state ORDER BY sales) as next_sales
FROM revenue;

state city sales next_sales

California 1 10000 10000
California 2 10000 40000
California 3 40000 30000
California 4 30000 Null
Texas 1 50000 50000
Texas 2 50000 60000
Texas 3 60000 60000
Texas 4 60000 Null
ach of the states
) as next_sales
10 Write a query to find the sales of the previous month in each of the states

Revenue
state month sales
California 4 30000
California 1 10000
California 3 40000
California 2 10000
Texas 1 50000
Texas 4 60000
Texas 3 60000
Texas 2 50000

Let's first partition the states and then sort the data by month

state month sales

California 1 10000
California 2 10000
California 3 40000
California 4 30000
Texas 1 50000
Texas 2 50000
Texas 3 60000
Texas 4 60000

Now, let's find the sales of the previous month

state month sales prev_sales

California 1 10000 Null
California 2 10000 10000
California 3 40000 10000
California 4 30000 40000
Texas 1 50000 Null
Texas 2 50000 50000
Texas 3 60000 50000
Texas 4 60000 60000

How can I write a query to get this answer programmatically?

 LAG(sales) OVER(PARTITION BY state ORDER BY sales) as prev_sales
FROM revenue;

state city sales prev_sales

California 1 10000 10000
California 2 10000 40000
California 3 40000 30000
California 4 30000 Null
Texas 1 50000 50000
Texas 2 50000 60000
Texas 3 60000 60000
Texas 4 60000 Null
ch of the states
as prev_sales
11 Write a query to find the very first sale in each of the states

Revenue
state month sales
California 4 30000
California 1 10000
California 3 40000
California 2 10000
Texas 1 50000
Texas 4 60000
Texas 3 60000
Texas 2 50000

Let's first partition the states and then sort the data by month

state month sales

California 1 10000
California 2 10000
California 3 40000
California 4 30000
Texas 1 50000
Texas 2 50000
Texas 3 60000
Texas 4 60000

Now, let's find the sales of the first month

state month sales first_sales

California 1 10000 10000
California 2 10000 10000
California 3 40000 10000
California 4 30000 10000
Texas 1 50000 50000
Texas 2 50000 50000
Texas 3 60000 50000
Texas 4 60000 50000

How can I write a query to get this answer programmatically?

 FIRST_VALUE(sales) OVER(PARTITION BY state ORDER BY sales) as first_sales
FROM revenue;

state city sales first_sales

California 1 10000 10000
California 2 10000 10000
California 3 40000 10000
California 4 30000 10000
Texas 1 50000 50000
Texas 2 50000 50000
Texas 3 60000 50000
Texas 4 60000 50000
BY sales) as first_sales
12 Write a query to find the very last sale in each of the states

Revenue
state month sales
California 4 30000
California 1 10000
California 3 40000
California 2 10000
Texas 1 50000
Texas 4 60000
Texas 3 60000
Texas 2 50000

Let's first partition the states and then sort the data by month

state month sales

California 1 10000
California 2 10000
California 3 40000
California 4 30000
Texas 1 50000
Texas 2 50000
Texas 3 60000
Texas 4 60000

Now, let's find the sales of the last month

state month sales last_sale

California 1 10000 30000
California 2 10000 30000
California 3 40000 30000
California 4 30000 30000
Texas 1 50000 60000
Texas 2 50000 60000
Texas 3 60000 60000
Texas 4 60000 60000

How can I write a query to get this answer programmatically?

 LAST_VALUE(sales) OVER(PARTITION BY state ORDER BY sales) as last_sale
FROM revenue;

state city sales last_sale

California 1 10000 30000
California 2 10000 30000
California 3 40000 30000
California 4 30000 30000
Texas 1 50000 60000
Texas 2 50000 60000
Texas 3 60000 60000
Texas 4 60000 60000
Y sales) as last_sale