SQL Window Functions Cheat Sheet

Window Functions Aggregate Functions vs. Window Functions PARTITION BY ORDER BY

compute their result based on a sliding window unlike aggregate functions, window functions do not collapse rows. divides rows into multiple groups, called partitions, to specifies the order of rows in each partition to which the
frame, a set of rows that are somehow related to which the window function is applied. window function is applied.
the current row. PARTITION BY city PARTITION BY city ORDER BY month
Aggregate Functions Window Functions month city sold month city sold sum sold city month sold city month
∑ 1 Rome 200 1 Paris 300 800 200 Rome 1 300 Paris 1
2 Paris 500 2 Paris 500 800 500 Paris 2 500 Paris 2
∑ 1 London 100 1 Rome 200 900 100 London 1 200 Rome 1
current row 1 Paris 300 2 Rome 300 900 300 Paris 1 300 Rome 2
∑ ∑ 2 Rome 300 3 Rome 400 900 300 Rome 2 400 Rome 3
2 London 400 1 London 100 500 400 London 2 100 London 1
3 Rome 400 2 London 400 500 400 Rome 3 400 London 2

Default Partition: with no PARTITION BY clause, the entire Default ORDER BY: with no ORDER BY clause, the order of
Syntax result set is the partition. rows within each partition is arbitrary.

SELECT city, month, SELECT <column_1>, <column_2>, Window Frame

sum(sold) OVER ( <window_function>() OVER ( is a set of rows that are somehow related to the current row. The window frame is evaluated separately within each partition.
PARTITION BY city PARTITION BY <...>
ORDER BY month ORDER BY <...> ROWS | RANGE | GROUPS BETWEEN lower_bound AND upper_bound
RANGE UNBOUNDED PRECEDING) total <window_frame>) <window_column_alias>
PARTITION UNBOUNDED
FROM sales; FROM <table_name>; PRECEDING The bounds can be any of the five options:
n PRECEDING
∙ 
UNBOUNDED PRECEDING
N ROWs
Named Window Definition ∙ 
n PRECEDING
CURRENT
ROW ∙ 
CURRENT ROW
M ROWs ∙ 
n FOLLOWING
SELECT country, city, SELECT <column_1>, <column_2>, M FOLLOWING ∙ 
UNBOUNDED FOLLOWING
rank() OVER country_sold_avg <window_function>() OVER <window_name> UNBOUNDED
FROM sales FROM <table_name> FOLLOWING The lower_bound must be BEFORE the upper_bound
WHERE month BETWEEN 1 AND 6 WHERE <...>
GROUP BY country, city GROUP BY <...>
ROWS BETWEEN 1 PRECEDING RANGE BETWEEN 1 PRECEDING GROUPS BETWEEN 1 PRECEDING
HAVING sum(sold) > 10000 HAVING <...> AND 1 FOLLOWING AND 1 FOLLOWING AND 1 FOLLOWING
WINDOW country_sold_avg AS ( WINDOW <window_name> AS ( city sold month city sold month city sold month
PARTITION BY country PARTITION BY <...> Paris 300 1 Paris 300 1 Paris 300 1
ORDER BY avg(sold) DESC) ORDER BY <...> Rome 200 1 Rome 200 1 Rome 200 1
ORDER BY country, city; <window_frame>) Paris 500 2 Paris 500 2 Paris 500 2
ORDER BY <...>; current
Rome 100 4
current
Rome 100 4
current
Rome 100 4
row Paris 200 4 row Paris 200 4 row Paris 200 4
Paris 300 5 Paris 300 5 Paris 300 5
Rome 200 5 Rome 200 5 Rome 200 5
PARTITION BY, ORDER BY, and window frame definition are all optional. London 200 5 London 200 5 London 200 5
London 100 6 London 100 6 London 100 6
Rome 300 6 Rome 300 6 Rome 300 6
Logical Order of Operations in SQL 1 row before the current row and values in the range between 3 and 5 1 group before the current row and 1 group
1 row after the current row ORDER BY must contain a single expression after the current row regardless of the value

1. FROM, JOIN 7. SELECT As of 2020, GROUPS is only supported in PostgreSQL 11 and up.
2. WHERE 8. DISTINCT
3. GROUP BY 9. UNION/INTERSECT/EXCEPT
4. aggregate functions 10. ORDER BY Abbreviations Default Window Frame
5. HAVING 11. OFFSET
6. window functions 12. LIMIT/FETCH/TOP Abbreviation Meaning If ORDER BY is specified, then the frame is
UNBOUNDED PRECEDING BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW RANGE BETWEEN UNBOUNDED PRECEDING AND
n PRECEDING BETWEEN n PRECEDING AND CURRENT ROW CURRENT ROW.
You can use window functions in SELECT and ORDER BY. However, you can’t put window functions anywhere in the FROM, CURRENT ROW BETWEEN CURRENT ROW AND CURRENT ROW Without ORDER BY, the frame specification is
WHERE, GROUP BY, or HAVING clauses. n FOLLOWING BETWEEN AND CURRENT ROW AND n FOLLOWING ROWS BETWEEN UNBOUNDED PRECEDING AND
UNBOUNDED FOLLOWING BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING UNBOUNDED FOLLOWING.

LearnSQL.com is owned by Vertabelo SA

Try out the interactive Window Functions course at LearnSQL.com, and check out our other SQL courses. vertabelo.com | CC BY-NC-ND Vertabelo SA
SQL Window Functions Cheat Sheet
List of Window Functions Ranking Functions Distribution Functions
∙ row_number() − unique number for each row within partition, with different numbers ∙ 
percent_rank() − the percentile ranking number of a row—a value in [0, 1] interval:
Aggregate Functions for tied values (rank - 1) / (total number of rows - 1)
∙ avg() ∙ rank() − ranking within partition, with gaps and same ranking for tied values ∙ 
cume_dist() − the cumulative distribution of a value within a group of values, i.e., the number of
∙ count() ∙ dense_rank() − ranking within partition, with no gaps and same ranking for tied values rows with values less than or equal to the current row’s value divided by the total number of rows;
∙ max() a value in (0, 1] interval
∙ 
min() row_number rank dense_rank
city price cume_dist() OVER(ORDER BY sold) percent_rank() OVER(ORDER BY sold)
over(order by price)
∙ sum() Paris 7 1 1 1 city sold cume_dist city sold percent_rank
Rome 7 2 1 1 Paris 100 0.2 Paris 100 0
Ranking Functions London 8.5 3 3 2 Berlin 150 0.4 Berlin 150 0.25
∙ row_number() Berlin 8.5 4 3 2 Rome 200 0.8 Rome 200 0.5
Moscow 9 5 5 3 80% of values are without this row 50% of
∙ rank() Moscow 200 0.8 Moscow 200 0.5
Madrid 10 6 6 4 less than or equal values are less than this
London 300 1 London 300 1
∙ dense_rank() Oslo 10 7 6 4 to this one row’s value

Distribution Functions ORDER BY and Window Frame: rank() and dense_rank() require ORDER BY, but ORDER BY and Window Frame: Distribution functions require ORDER BY. They do not accept window frame
∙ percent_rank() row_number() does not require ORDER BY. Ranking functions do not accept window definition (ROWS, RANGE, GROUPS).
frame definition (ROWS, RANGE, GROUPS).
∙ cume_dist()
Analytic Functions
∙ lead()
Analytic Functions ∙ 
first_value(expr) − the value for the first row within the window frame
∙ 
lead(expr, offset, default) − the value for the row offset rows after the current; offset and ∙ 
last_value(expr) − the value for the last row within the window frame
∙ lag()
default are optional; default values: offset = 1, default = NULL
∙ 
ntile()
∙ 
lag(expr, offset, default) − the value for the row offset rows before the current; offset and first_value(sold) OVER last_value(sold) OVER
∙ first_value()
default are optional; default values: offset = 1, default = NULL (PARTITION BY city ORDER BY month) (PARTITION BY city ORDER BY month
∙ last_value() RANGE BETWEEN UNBOUNDED PRECEDING
lag(sold) OVER(ORDER BY month) lead(sold) OVER(ORDER BY month) city month sold first_value AND UNBOUNDED FOLLOWING)
∙ nth_value() Paris 1 500 500
month sold month sold city month sold last_value
order by month

order by month
Paris 2 300 500
1 500 NULL 1 500 300 Paris 3 400 500 Paris 1 500 400
2 300 500 2 300 400 Rome 2 200 200 Paris 2 300 400
Paris 3 400 400
Aggregate Functions 3 400 300 3 400 100 Rome 3 300 200
Rome 2 200 500
4 100 400 4 100 500 Rome 4 500 200
5 500 100 5 500 NULL Rome 3 300 500
∙ 
avg(expr) − average value for Rome 4 500 500
rows within the window frame
lag(sold, 2, 0) OVER(ORDER BY month) lead(sold, 2, 0) OVER(ORDER BY month)
Note: You usually want to use RANGE BETWEEN
month sold month sold
order by month

order by month

∙ 
count(expr) − count of values UNBOUNDED PRECEDING AND UNBOUNDED
offset=2

1 500 0 1 500 400

for rows within the window FOLLOWING with last_value(). With the default
2 300 0 2 300 100
frame 3 400 500 3 400 500 window frame for ORDER BY, RANGE UNBOUNDED

offset=2
4 100 300 4 100 0 PRECEDING, last_value() returns the value for
∙ 
m ax(expr) − maximum value
 5 500 400 5 500 0 the current row.
within the window frame

∙ 
ntile(n) − divide rows within a partition as equally as possible into n groups, and assign each ∙ 
nth_value(expr, n) − the value for the n-th row within the window frame; n must be an integer
∙ 
min(expr) − minimum value
row its group number.
within the window frame nth_value(sold, 2) OVER
ntile(3) (PARTITION BY city ORDER BY month)
∙ 
sum(expr) − sum of values within city sold city month sold nth_value
the window frame Rome 100 1 Paris 1 500 300
Paris 100 1 1 1 Paris 2 300 300
London 200 1 Paris 3 400 300
Moscow 200 2 Rome 2 200 300
ORDER BY and Window Frame: Berlin 200 2 2 2
ORDER BY and Window Frame: ntile(), ORDER BY and Window Frame: first_value(),
Rome 3 300 300
Aggregate functions do not require an Madrid 300 2 lead(), and lag() require an ORDER BY. Rome 4 500 300
last_value(), and nth_value() do not
ORDER BY. They accept window frame Oslo 300 3 They do not accept window frame definition Rome 5 300 300 require an ORDER BY. They accept window frame
3 3
definition (ROWS, RANGE, GROUPS). Dublin 300 3 (ROWS, RANGE, GROUPS). London 1 100 NULL definition (ROWS, RANGE, GROUPS).

LearnSQL.com is owned by Vertabelo SA

Try out the interactive Window Functions course at LearnSQL.com, and check out our other SQL courses. vertabelo.com | CC BY-NC-ND Vertabelo SA