Window Function Nedir?
Similar to an aggregate function, a window function operates on a set of rows. However, it does not reduce the number of rows returned by the query.
The term window describes the set of rows on which the window function operates. A window function returns values from the rows in a window.
Söz dizimi
şöyle. Yani aslında bir şeyleri grupluyor ve gruptaki her satır için bir şey hesaplıyor
window_function(arg1, arg2,..) OVER (
[PARTITION BY partition_expression]
[ORDER BY sort_expression [ASC | DESC] [NULLS {FIRST | LAST }])
Eğer
PARTITION BY kullanılmazsa tek grup varmış gibi davranır. Gruptaki satırlar ORDER BY ile sıralanabilir.
Window Function Nedir vs SubSelect/Subquery
In a standards-compliant RDBMS, you can eliminate some of your subqueries and your in clause by using a windowing expression:
Örnek
Elimizde şöyle bir kod
olsun. Harcama toplamı belli bir miktarı geçen müşterileri bulmak isteyelim. Burada GROUP BY + HAVING ile harcama tablosundaki müşteriler seçiliyor. Daha sonra JOIN ile müşteri tablosu ile birleştiriliyor.
SELECT exp.employee_id, emp.name, exp.amount
FROM Employee emp
JOIN Expenditure exp ON exp.employee_id = emp.id
WHERE exp.employee_id IN (
SELECT employee_id
FROM Expenditure
GROUP BY employee_id
HAVING SUM(amount) >= 100::money
)
ORDER BY exp.amount DESC;SELECT exp.employee_id, emp.name, exp.amount
FROM Employee emp
JOIN (
SELECT
employee_id, amount,
SUM(amount) OVER (PARTITION BY employee_id) AS total
FROM Expenditure
) exp ON exp.employee_id = emp.id
WHERE exp.total >= 100::money
ORDER BY exp.amount DESC;frame_clause Nedir?
Sorguda birden fazla Window Function kullanabilmeyi sağlar.
Örnek
SELECT
wf1() OVER(PARTITION BY c1 ORDER BY c2),
wf2() OVER(PARTITION BY c1 ORDER BY c2)
FROM table_name;Eğer sürekli PARTITION BY yazmak istemiyorsak bunu kısaltmak için şöyle
yaparız. Burada WINDOW kullanılıyor
SELECT
wf1() OVER w,
wf2() OVER w,
FROM table_name
WINDOW w AS (PARTITION BY c1 ORDER BY c2);
Window Function Listesi
Bir açıklamaya göre Window Function Listesi
şöyle. Bence bu listeye AVG ve diğer bazı şeyler de eklenmeli
Name Description
CUME_DIST Return the relative rank of the current row.
DENSE_RANK Rank the current row within its partition without gaps.
FIRST_VALUE Return a value evaluated against the first row within its partition.
LAG Return a value evaluated at the row that is at a specified physical offset row before the current row within the partition.
LAST_VALUE Return a value evaluated against the last row within its partition.
LEAD Return a value evaluated at the row that is offset rows after the current row within the partition.
NTILE Divide rows in a partition as equally as possible and assign each row an integer starting from 1 to the argument value.
NTH_VALUE Return a value evaluated against the nth row in an ordered partition.
PERCENT_RANK Return the relative rank of the current row (rank-1) / (total rows – 1)
RANK Rank the current row within its partition with gaps.
ROW_NUMBER Number the current row within its partition starting from 1.
Başka bir açıklamaya göre Basic Window Functions
şöyle. Neyse kimin ne dediği önemli değil
Aggregate Functions : SUM, COUNT, MIN, MAX, AVG
Ranking Functions : RANK, DENSE_RANK, ROW_NUMBER, NTILE
Value Functions : LAG, LEAD, FIRST_VALUE, LAST_VALUE
FIRST_VALUE
Elimizde şöyle bir tablo
olsun. Her sütundaki
son null
olmayan değeri isteyelim. Bu tablodaki sonuç (A1,B2,C1) olmalı
version col_A col_B col_C
1 A1 B1 (null)
2 A2 B3 (null)
3 A3 B2 (null)
4 A5 (null) C1
5 A1 (null) (null)
Şöyle
yaparız. Burada frame_clause kullanılıyor. Sorguyu kısaltmak için WINDOW da kullanılabilirdi. ORDER BY ile NULL değer 0 ile değiştiriliyor ve azalan şekilde sıralanıyor. Böylece FIRST_VALUE en son sütun değerini buluyor
SELECT DISTINCT
FIRST_VALUE(col_A) OVER (ORDER BY CASE WHEN col_A IS NULL THEN 0 ELSE version END DESC)
AS LastA,
FIRST_VALUE(col_B) OVER (ORDER BY CASE WHEN col_B IS NULL THEN 0 ELSE version END DESC)
AS LastB,
FIRST_VALUE(col_C) OVER (ORDER BY CASE WHEN col_C IS NULL THEN 0 ELSE version END DESC)
AS LastC
FROM tab
LAST_VALUE
Q : Write a SQL query to find the total number of people who are inside the hospital.
A : the last_value window function is used to find the last value of a particular partition.
So here we can use this to find the last action of each user.
And if the last action is 'in', then the user is inside the hospital.
SELECT cte AS (
SELECT*,LAST_VALUE(action) OVER(PARTITION BY emp_id) = 'in' AS 'in_action'
FROM hospital
)
SELECT DISTINCT emp_id FROM cte WHERE
in_action = 1
