Question

我有一个SQL Server表和一个查询，它会产生一个带有两个数据透视列的数据透视表。以前它只使用一个枢轴列（pvt1）。

SELECT
    MIN([Created Date]),
    MIN([2]) AS [2],
    MIN([4]) AS [4],
    MIN([12]) AS [12],
    MIN([18]) AS [18],
    MIN([47]) AS [47],
    MIN([88]) AS [88],
    MIN([982]) AS [982],
    MIN([1033]) AS [1033],
    MIN([2360]) AS [2360],
    MIN([3759]) AS [3759],
    MIN([3974]) AS [3974],
    MIN([5586]) AS [5586]
FROM 
    (SELECT
         year(udp.CreatedDate) AS "Created Date",
         MAX(udp.Participant) AS "Participant",
         MAX(udp.LastModifiedDate) AS "Last Modified Date",
         MAX(udp.Client) AS "Client"
     FROM
         UDP_Table AS udp
     WHERE
         1=1
     GROUP BY
         year(udp.CreatedDate)) AS source
PIVOT (
    MAX([Last Modified Date])
FOR
  [Participant] IN ([2], [4], [12], [18], [47], [88])
)
as pvt1
PIVOT (
    MAX([Last Modified Date])
FOR
   [Client] IN ([982], [1033], [2360], [3759], [3974], [5586])
)
as pvt2

GROUP BY
    [Created Date]
ORDER BY
    [Created Date] asc

结果是：

[S0001] [207]无效的列名称“上次修改日期”。

奇怪的是当我尝试这个示例表和查询（参见下面的链接）时，它具有与我相同的结构，它可以工作。

http://pratchev.blogspot.de/2009/01/pivoting-on-multiple-columns.html

我的查询有什么问题？

Answer 1

您的第一个数据透视表将列[Last Modified Date]转换为名为[2], [4], [12], [18], [47], [88]的列，因此[Last Modified Date]不存在第二个数据透视表。您发布的示例为每个分支SUM(value) AS value和SUM(quantity) AS quantity

使用了2个不同的列

我不是100％确定您要完成的任务，但根据您的最终SELECT，您需要为每个参与者和每个客户创建一个列。你应该做一个UNION来加入记录，然后只做1个Pivot。

这看起来像这样。

SELECT *
FROM 
    (SELECT
         year(udp.CreatedDate) AS "Created Date",
         MAX(udp.Participant) AS "Participant",
         MAX(udp.LastModifiedDate) AS "Last Modified Date",
     FROM
         UDP_Table AS udp
     GROUP BY
         year(udp.CreatedDate)
     UNION ALL
     SELECT
         year(udp.CreatedDate) AS "Created Date",
         MAX(udp.Client) AS "Participant",
         MAX(udp.LastModifiedDate) AS "Last Modified Date",
     FROM
         UDP_Table AS udp
     GROUP BY
         year(udp.CreatedDate)
) AS source
PIVOT (
    MAX([Last Modified Date])
    FOR [Participant] IN ([2], [4], [12], [18], [47], [88], [982], [1033], [2360], [3759], [3974], [5586])
)
as pvt1

T-SQL多重轴支柱问题

1 个答案: