SQL Server 2008 R2：ROW_NUMBER（）有2列

Question

表結構：

CREATE TABLE AZTool
(
    t_ID int,
    z_ID int,
    col_date date
);

插入數據：

INSERT INTO AZTool values(12,23409,'2017-01-02')
INSERT INTO AZTool values(12,23409,'2017-01-03')
INSERT INTO AZTool values(21,23409,'2017-03-14')
INSERT INTO AZTool values(12,24455,'2017-04-22')
INSERT INTO AZTool values(22,24455,'2017-05-13')
INSERT INTO AZTool values(22,35600,'2017-04-04')
INSERT INTO AZTool values(23,35600,'2017-05-14')
INSERT INTO AZTool values(24,35600,'2017-05-16')
INSERT INTO AZTool values(25,35600,'2017-05-24')

預期產量：

t_ID    z_ID    RowNumber
-------------------------
12      23409   1
12      23409   1
21      23409   2
12      24455   1
22      24455   2
22      35600   1
23      35600   2
24      35600   3
25      35600   4

我的嘗試：

SELECT
    t_ID, z_ID,
    ROW_NUMBER() OVER (PARTITION BY t_ID, z_ID ORDER BY z_ID) rn
FROM 
    AZTool

Answer 1

我糾正的問題：

• 從分區中刪除t_ID
• 在ORDER BY子句中按t_ID
• 使用DENSE_RANK而不是ROW_NUMBER

SELECT
    t_ID,
    z_ID,
    DENSE_RANK() OVER(PARTITION BY z_ID ORDER BY t_ID) rn
FROM AZTool;

Answer 2

使用DENSE_RANK()函數：

SELECT t_ID,z_ID,DENSE_RANK() OVER(PARTITION BY z_ID ORDER BY t_ID) rn
FROM AZTool

結果：