在SQL Server 2008 R2中使用PIVOT表

Question

我遇到了一个我无法解决的问题，已经尝试了一切。希望我能在这里得到答案。

我表中的数据如下

Id Month Val1 Val2 Val3
1  Jan   70   80   90
2  Jan   12   13   15
3  Feb   12   67   99
4  March 14   15   17.1

从以下结构中，我想基于列透视数据，val1这样我的输出将类似于：

Jan Jan Feb March
70  12  12  14

Answer 1

我认为您将需要使用动态 SQL（除非您的列保持静态）

需要执行的查询是：

SELECT  [Jan] = [Jan1], 
        [Jan] = [Jan2], 
        [Feb] = [Feb1], 
        [March] = [March1]
FROM    (   SELECT  [PivotColumn] = [Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID)), 
                    Val1
            FROM    T
        ) t
        PIVOT
        (   MAX(Val1)
            FOR [PivotColumn] IN ([Jan1], [Jan2], [Feb1], [March1])
        ) pvt

虽然我不确定为什么会这样

Jan Jan Feb March
70  12  12  14

并不是

Jan Jan Feb March
12  70  12  14

所以你可能需要修改函数ORDER BY中的ROW_NUMBER。

要动态构建它，您可以使用：

-- CREATE SAMPLE TABLE AND INSERT DATA
CREATE TABLE #T (ID INT, Month VARCHAR(10), Val1 INT, Val2 INT, Val3 DECIMAL(5, 1));
INSERT #T VALUES (1, 'Jan', 70, 80, 90), (2, 'Jan', 12, 13, 15), (3, 'Feb', 12, 67, 99), (4, 'March', 14, 15, 17.1);

-- DECLARE VARIABLES TO STORE THE COLUMN NAMES
DECLARE @PivotList NVARCHAR(MAX) = '',
        @ColumnList NVARCHAR(MAX) = '';

-- HERE USE ROW_NUMBER() TO UNIQUELY IDENTIFY VALUES FOR MONTHS 
-- THIS MEANS JAN: 70 AND JAN: 12 CAN BE IDENTIFIED SEPARATELY LATER, BUT RETAIN THE DUPLICATE COLUMN NAME [Jan]
SELECT  @ColumnList = @ColumnList + ', ' + QUOTENAME([Month]) + ' = ' + QUOTENAME([Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID))),
        @PivotList = @PivotList + ', ' + QUOTENAME([Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID)))
FROM    #T
ORDER BY ID;

DECLARE @SQL NVARCHAR(MAX) = 'SELECT ' + STUFF(@ColumnList, 1, 2, '') + '
                            FROM    (   SELECT  [PivotColumn] = [Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID)), 
                                                Val1
                                        FROM    #T
                                    ) t
                                    PIVOT
                                    (   MAX(Val1)
                                        FOR [PivotColumn] IN (' + STUFF(@PivotList, 1, 2, '') + ')
                                    ) pvt';

EXECUTE SP_EXECUTESQL @SQL;

DROP TABLE #T;

同样，意志中的任何更改ROW_NUMBER也需要反映在ORDER BY生成列名称和数据透视表的查询中的子句中：

SELECT  @ColumnList = @ColumnList + ', ' + QUOTENAME([Month]) + ' = ' + QUOTENAME([Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID))),
        @PivotList = @PivotList + ', ' + QUOTENAME([Month] + CONVERT(VARCHAR(10), ROW_NUMBER() OVER(PARTITION BY [Month] ORDER BY ID)))
FROM    #T
ORDER BY ID;