具有多个列和动态列名称的Unpivot表

时间:2012-08-09 10:15:19

标签: sql dynamic multiple-columns unpivot dynamic-columns

我正在尝试将具有多行和多列的表取消。需要将每行扩展为包含特定列的2行,并且需要重命名列名称,并根据所选列添加新列!

我包括样本数据之前和之后以及设置数据的脚本。

 CREATE TABLE #tmpProducts (
        ProductId INT,
        ProductName nVARCHAR(100),
        B2B_GrossRevenue DECIMAL(10,2),
        B2B_DirectCost DECIMAL(10,2),
        B2B_NetRevenue DECIMAL(10,2),
        B2C_GrossRevenue DECIMAL(10,2),
        B2C_DirectCost DECIMAL(10,2),
        B2C_NetRevenue DECIMAL(10,2)
    )

    INSERT INTO #tmpProducts SELECT 1, 'Product1',1545.00,406.25,1138.75,195.00,35.10,159.90
    INSERT INTO #tmpProducts SELECT 2, 'Product2',902.00,189.00,713.00,3280.00,590.40,2689.60
    INSERT INTO #tmpProducts SELECT 3, 'Product3',15665.00,3988.39,11676.61,6247.00,1124.46,5122.54
    INSERT INTO #tmpProducts SELECT 4, 'Product4',736.00,196.16,539.84,2395.00,431.10,1963.90

    SELECT * FROM #tmpProducts
    DROP TABLE #tmpProducts

    CREATE TABLE #tmpProducts2 (
        ProductId INT,
        ProductName nVARCHAR(100),
        [Type] nVARCHAR(3),
        GrossRevenue DECIMAL(10,2),
        DirectCost DECIMAL(10,2),
        NetRevenue DECIMAL(10,2)
    )

    INSERT INTO #tmpProducts2 SELECT 1, 'Product1','B2B',1545.00,406.25,1138.75
    INSERT INTO #tmpProducts2 SELECT 1, 'Product1','B2C',195.00,35.10,159.90
    INSERT INTO #tmpProducts2 SELECT 2, 'Product2','B2B',902.00,189.00,713.00
    INSERT INTO #tmpProducts2 SELECT 2, 'Product2','B2C',3280.00,590.40,2689.60
    INSERT INTO #tmpProducts2 SELECT 3, 'Product3','B2B',15665.00,3988.39,11676.61
    INSERT INTO #tmpProducts2 SELECT 3, 'Product3','B2C',6247.00,1124.46,5122.54
    INSERT INTO #tmpProducts2 SELECT 4, 'Product4','B2B',736.00,196.16,539.84
    INSERT INTO #tmpProducts2 SELECT 4, 'Product4','B2C',2395.00,431.10,1963.90

    SELECT * FROM #tmpProducts2
    DROP TABLE #tmpProducts2

我试过这个,但我无法通过第二列,我不知道如何添加一个具有特定文本的新列,(可能是动态sql,但试图避免这种情况,如果可能的话)

这是我尝试的开始,非常感谢任何帮助。

SELECT  ProductId, ProductName,GrossRevenue
        FROM    (
                    SELECT ProductId, ProductName, B2B_GrossRevenue,B2C_GrossRevenue FROM #tmpProducts
                ) as t
            UNPIVOT ( GrossRevenue for test IN (B2B_GrossRevenue,B2C_GrossRevenue)) AS unpvt

3 个答案:

答案 0 :(得分:5)

你只是保持不动摇

    SELECT  ProductId,ProductName, Substring(col1,1,3) as type,  GrossRevenue, DirectCost, NetRevenue
    FROM    ( 
        SELECT * FROM #tmpProducts 
    ) as t 
    UNPIVOT ( GrossRevenue for col1 IN (B2B_GrossRevenue,B2C_GrossRevenue)) AS unpvt 
    unpivot ( DirectCost for col2 in (B2b_DirectCost, B2c_DirectCost)) up2
    unpivot ( NetRevenue for col3 in (B2b_NetRevenue, B2c_NetRevenue)) up3
where SUBSTRING(col1,1,3)=SUBSTRING(col2,1,3)
and  SUBSTRING(col1,1,3)=SUBSTRING(col3,1,3)

并加入col列以过滤掉不匹配

答案 1 :(得分:2)

以下是使用Dynamic SQL执行此操作的方法。这将允许您在执行时获取所有列。然后,如果您获得任何数据更改,则无需更改查询。这同时适用于UNPIVOTPIVOT

DECLARE @colsUnPivot AS NVARCHAR(MAX),
    @colsPivot AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

SET @colsUnPivot = stuff((select ','+quotename(C.name)
         from sys.columns as C
         where C.object_id = object_id('tmpProducts') and
               C.name like 'B2%'
         for xml path('')), 1, 1, '')

SET @colsPivot = stuff((select DISTINCT ','+quotename(right((C.name), len(C.name)-4))
         from sys.columns as C
         where C.object_id = object_id('tmpProducts') and
               C.name like 'B2%'
         for xml path('')), 1, 1, '')


set @query 
  = ' SELECT ProductId, ProductName, [type], ' + @colsPivot +'
      FROM
      (
        SELECT ProductId, ProductName, substring(field, 1, 3) [type]
                ,  value, right((field), len(field)-4) as col
        from tmpProducts
        unpivot 
        (
           value
           for field in (' + @colsUnPivot + ')
        ) unpvt 
      ) x
      PIVOT
      (
        sum(value)
        FOR col IN (' + @colsPivot +')
      )p'


execute(@query)

请参阅SQL Fiddle with Demo

答案 2 :(得分:0)

如果您有多个想要转移的群组,您也可以这样做:

Select *
  from table
  unpivot include nulls ( (test_name, test_date) for col_nm in ( (test1, date1) as 'test_val1'
                                                                ,(test2, date2) as 'test_val2'
                                                                ,(test3, date3) as 'test_val3'
                                                               )
                     )