Question

我有下表，我需要总结一下

ID      A   B   C   D   E   F   G
----------------------------------
1-100   1   2   1   1   1   1   1
1-201   1   2   1   2   2   2   2
1-322   1   1   1   1   2   2   1
2-155   1   1   2   1   1   2   2
2-167   2   1   2   1   2   1   2
2-389   2   2   1   2   1   1   2
2-423   1   2   2   2   1   1   1
3-10    2   1   1   1   2   2   2
3-222   1   1   1   1   2   2   1
3-397   2   1   1   2   2   1   1

在上表中，值1编码为S，而2编码为R.此外，ID是代码XX，YY或XX，其中 - 之前的数字代表XX，YY或XX。 / p>

我想要的摘要是这个

         XX         YY              ZZ
------------------------------------------
A   S   3   100%    2   50%     1   33%
    R   0   0%      2   50%     2   66%
B   S   2   66%     2   50%     3   100%
    R   1   33%     2   50%     0   0%
C   S   3   100%    3   75%     3   100%
    R   0   0%      1   25%     0   0%
D   S   2   66%     2   50%     2   66%
    R   1   33%     2   50%     1   33%
E   S   1   33%     3   75%     0   0%
    R   2   66%     1   25%     3   100%
F   S   1   33%     3   75%     2   66%
    R   2   66%     1   25%     1   33%
G   S   2   66%     1   25%     1   33%
    R   1   33%     3   75%     2   66%

所以我需要旋转表，计算1/2并创建百分比。

这让我感到非常困惑，我已经做了一些关于如何做到这一点的死胡同（更不用说如何优雅地做到这一点）

提前致谢！

在马丁的帮助下，我非常亲近。我的数据当然比我给出的例子有点愚蠢，所以我仍然遇到困难。我已经对数据进行了谴责，并将正确的编码放入我想要的内容中 - 是的，编码是真正的傻瓜，我无法控制它们。）

我已将Martins SQL扩展为链接到我的数据，但还有两个问题。 Thing Column中行的顺序不是我想要的。

当我尝试以下代码时，我得到一个“必须声明标量变量@order” - 它不喜欢加入我的名为myOrder的临时表。

DECLARE @myOrder TABLE (rug varchar(3), rugOrder int)    
INSERT @myOrder
    SELECT 'INH', 1 UNION ALL
    SELECT 'RIF', 2 UNION ALL
    SELECT 'KM', 3 UNION ALL
    SELECT 'AK', 4 UNION ALL
    SELECT 'CM', 5 UNION ALL
    SELECT 'MOX', 6 UNION ALL
    SELECT 'OFX', 7;

WITH YourData(ID, INH, RIF, KM, AK, CM, MOX, OFX) As
(SELECT Sample_ID, INH, RIF, KM, AK, CM, MOX, OFX
FROM dbo.[GCT_Rug] WHERE Sample_ID NOT LIKE '99%')

, Unpivoted AS
(
SELECT S_R_Flag,
       Thing,
       Site = 
        CASE 
        WHEN LEFT(ID,1) = 1 THEN 1 
        WHEN LEFT(ID,1) = 6 THEN 1 
        WHEN LEFT(ID,1) = 8 THEN 2 
        WHEN LEFT(ID,1) = 9 THEN 3 END

FROM YourData
UNPIVOT
   (S_R_Flag FOR Thing IN (INH, RIF, KM, AK, CM, MOX, OFX)
)AS unpvt)
SELECT Thing
       ,SRFLAG =
            CASE 
                WHEN S_R_Flag = 1 THEN 'S'
                WHEN S_R_Flag = 2 THEN 'R'
            END
       ,[1] AS IND
   ,round(CAST([1] AS FLOAT) / NULLIF(SUM([1]) OVER (PARTITION BY Thing),0)*100,1) AS 'Ind Percent'
       ,[2] AS MD
   ,round(CAST([2] AS FLOAT) / NULLIF(SUM([2]) OVER (PARTITION BY Thing),0)*100,1) AS 'MD Percent'
   ,[3] AS 'SA'
   ,round(CAST([3] AS FLOAT) / NULLIF(SUM([3]) OVER (PARTITION BY Thing),0)*100,1) AS 'SA Percent'

FROM Unpivoted 
INNER JOIN @myOrder
ON Unpivoted.Thing= @myOrder.rug
PIVOT (COUNT (Site) FOR Site IN ( [1], [2], [3])) AS pvt
ORDER BY rugOrder,
         SRFLAG;

错误“必须声明标量变量@myOrder”是什么意思，为什么我不能加入呢？

再次感谢你们（特别是马丁）真棒！

Answer 1

这实际上为您提供了所需的结果，尽管我没有费心将数值映射到代码

;WITH YourData(ID,A,B,C,D,E,F,G) As
(
SELECT '1-100',1,2,1,1,1,1,1 UNION ALL
SELECT '1-201',1,2,1,2,2,2,2 UNION ALL
SELECT '1-322',1,1,1,1,2,2,1 UNION ALL
SELECT '2-155',1,1,2,1,1,2,2 UNION ALL
SELECT '2-167',2,1,2,1,2,1,2 UNION ALL
SELECT '2-389',2,2,1,2,1,1,2 UNION ALL
SELECT '2-423',1,2,2,2,1,1,1 UNION ALL
SELECT '3-10 ',2,1,1,1,2,2,2 UNION ALL
SELECT '3-222',1,1,1,1,2,2,1 UNION ALL
SELECT '3-397',2,1,1,2,2,1,1
), Unpivoted AS
(
SELECT S_R_Flag,
       Thing,
       SUBSTRING(ID,1,CHARINDEX('-',ID )-1) AS Code,ID
FROM YourData
UNPIVOT
   (S_R_Flag FOR Thing IN (A,B,C,D,E,F,G)
)AS unpvt)
SELECT Thing
       ,S_R_Flag
       ,[1]
       ,CAST([1] AS FLOAT) / SUM([1]) OVER (PARTITION BY Thing)
       ,[2]
       ,CAST([2] AS FLOAT) / SUM([2]) OVER (PARTITION BY Thing)
       ,[3]
       ,CAST([3] AS FLOAT) / SUM([3]) OVER (PARTITION BY Thing)
FROM Unpivoted
PIVOT (COUNT (ID) FOR Code IN ( [1], [2], [3] )) AS pvt
ORDER BY Thing,
         S_R_Flag;

T-SQL Count问题

1 个答案: