如何在动态数据透视中将两列组合为单列

Question

我有这样的架构

demo(id, val, month, year, decide)

演示数据和模式在这个小提琴http://sqlfiddle.com/#!3/dd89d5/1

中给出

在输出中，我想将行转换为列。 在我想要的输出中，

ID    (11 14)  (12 14) (2 15) ...     decider
101    0.45     0.5      0.3          411
102    0.4      0.2      0.1          411

我希望将月份，年份分类

我经历了很多线索，发现了一个小提琴。并尝试根据我的需要进行定制

以下是我尝试http://sqlfiddle.com/#!3/dd89d5/1

的内容

但无法得到解决方案。请帮帮我

Answer 1

这是你的表

create table demo(
    id varchar(max),    val decimal(4,2),    month int,    year int, decider int
  )

INSERT INTO demo
    ([id], [val], [month], [year], [decider])
VALUES

(101, 0.25, 11, 14, 411),
(101, 1, 12, 14, 411),
(101, 0.5, 1, 15, 411),
(101, 0.75, 2, 15, 411),
(102, 0.25, 11, 14, 411),
(102, 0.5, 12, 14, 411),
(102, 0.25, 1, 15, 411),
(101, 0.75, 11, 14, 412),
(101, 0.5, 1, 15, 412),
(101, 0.25, 2, 15, 412),
(102, 0.5, 11, 14, 412),
(102, 0.5, 12, 14, 412),
(103, 0.25, 1, 15, 412),
(103, 0.5, 11, 14, 411)
;

使用Dense_Rank订购数据透视列

SELECT DENSE_RANK() OVER(ORDER BY [YEAR] ,[MONTH]) RNO,*,CAST([MONTH] AS VARCHAR) + ' ' +  CAST([YEAR] AS VARCHAR) DT
INTO #TEMP
FROM
(
    SELECT ID,SUM(VAL)VAL,[MONTH],[YEAR],DECIDER
    FROM DEMO 
    GROUP BY ID,[MONTH],[YEAR],DECIDER
)TAB

选择pivot的列并声明一个变量以将NULL替换为零

DECLARE @cols NVARCHAR (MAX)
DECLARE @NullToZeroCols NVARCHAR (MAX)

SELECT @cols = COALESCE (@cols + ',[' + DT + ']', 
              '[' + DT + ']')
               FROM    (SELECT DISTINCT RNO,DT FROM #TEMP) PV  
               ORDER BY RNO

SET @NullToZeroCols = SUBSTRING((SELECT ',ISNULL(['+DT+'],0) AS ['+DT+']' 
FROM(SELECT DISTINCT DT,RNO FROM #TEMP GROUP BY DT,RNO)TAB  
ORDER BY RNO  FOR XML PATH('')),2,8000)

现在转动它

DECLARE @query NVARCHAR(MAX)
SET @query = 'SELECT P.ID,' + @NullToZeroCols + ',DECIDER FROM 
             (
                 SELECT ID, DT, val,DECIDER FROM #TEMP
             ) x
             PIVOT 
             (
                 SUM(val)
                 FOR DT IN (' + @cols + ')
            ) p
            ORDER BY ID;' 

EXEC SP_EXECUTESQL @query

结果

这是小提琴http://sqlfiddle.com/#!3/95111/1（如果在加载时出现任何错误，请按RUNSQL）