我在Sqlserver

Question

我有一个包含3列的表格：

A         B         C
test1     10
test1     15
test1     15
test2     1
test2     4
test2     3
test3     3
test3     3
test3     3

在第3栏中，我需要Count of Max(B) Row grouped on A。输出应该是：

A         B         C
test1     10        2
test1     15        2
test1     15        2
test2     1         1
test2     4         1
test2     3         1
test3     3         3
test3     3         3
test3     3         3

Answer 1

这是一个解决方案：

DECLARE @t TABLE ( A VARCHAR(20), B INT, C INT )

INSERT  INTO @t
        ( A, B )
VALUES  ( 'test1', 10 ),
        ( 'test1', 15 ),
        ( 'test1', 15 ),
        ( 'test2', 1 ),
        ( 'test2', 4 ),
        ( 'test2', 3 ),
        ( 'test3', 3 ),
        ( 'test3', 3 ),
        ( 'test3', 3 )

UPDATE  t1
SET     C = ca.C
FROM    @t t1
        CROSS APPLY ( SELECT TOP 1
                                COUNT(*) AS C
                      FROM      @t t2
                      WHERE     t2.A = t1.A
                      GROUP BY  B
                      ORDER BY  C DESC
                    ) ca      

SELECT  *
FROM    @t

输出：

A       B   C
test1   10  2
test1   15  2
test1   15  2
test2   1   1
test2   4   1
test2   3   1
test3   3   3
test3   3   3
test3   3   3

Answer 2

。 你好，

这可能是一个比已经发布的解决方案更好的解决方案，因为它只需要1/3的cpu power / ram，因为它可以批量生成所有内容，而不是间接循环请求。

-- Generate demo table
CREATE TABLE #temp(a nvarchar(20), b int, c int)

INSERT INTO #temp(a, b)
VALUES  (N'test1',10),
        (N'test1',15),
        (N'test1',15),
        (N'test2',1),
        (N'test2',4),
        (N'test2',3),
        (N'test3',3),
        (N'test3',3),
        (N'test3',3)
GO

-- Selecting the Result
SELECT t.a, t.b, mVal.c
FROM #temp as t
INNER JOIN (
    SELECT gmax.a, COUNT(gmax.b) as c
    FROM (
            SELECT t.a, t.b, MAX(b) OVER(PARTITION BY t.a) as maxVal -- Get Maximum Value per group
            FROM #temp as t
        ) as gmax
    WHERE gmax.b = gmax.maxVal -- get only the necessary rows
    GROUP BY gmax.a
) mVal
        ON t.a = mVal.a


-- Updating into #temp
UPDATE t
SET c = mVal.c
FROM #temp t
INNER JOIN (
    SELECT gmax.a, COUNT(gmax.b) as c
    FROM (
            SELECT t.a, t.b, MAX(b) OVER(PARTITION BY t.a) as maxVal 
            FROM #temp as t
        ) as gmax
    WHERE gmax.b = gmax.maxVal
    GROUP BY gmax.a
) mVal
    ON t.a = mVal.a

SELECT  *
FROM    #temp

DROP TABLE #temp
GO

只有Select和更新才对您有用。其余的仅用于演示目的。

祝你好运， 离子

Answer 3

我的方法是使用表格select distinct A and the max(B)中的@A加入@B group by A，A. B. test 1. 15 test 2. 4 test 3. 3 这样会产生以下结果：

set C =( select the count of A from your table  where A= @A  ) where A= @A

虽然Fitch的数据是数据循环 您可以更新表格

state = fields.Selection( [('new','New'), ('open','In Progress'), ('closed','Closed')], "Status", )

我跳了你的想法。