SQL Server支持2（或多个）聚合

Question

是否可以在SQL Server中的多个聚合列上进行透视？我有以下订单表：

[ORDERS]
| CustName | OrderedProduct | QtyOrdered | UnitCost | UnitPrice |  OrderDate
  Bob      | Canned Tuna    |          6 |       11 |        14 | 21-12-2016 13:11:00
  Steve    | Canned Salmon  |          2 |       15 |        19 | 03-11-2016 11:03:00
  Bob      | Canned Tuna    |         10 |       10 |        13 | 22-12-2016 10:43:00
  Bob      | Canned Tuna    |          4 |        9 |        10 | 13-11-2016 17:22:00

我有以下透视查询，它为我提供了11月/ 12月的每个客户的每件产品的订单数量：

SELECT CustName, OrderedProduct, [VOL Dec 2016], [VOL Nov 2016] FROM
(
    SELECT CustName, OrderedProduct, QtyOrdered,
    'VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthVol
    FROM [ORDERS]
) src
PIVOT
(
    SUM(QtyOrdered)
    FOR YearMonthVol IN ([VOL Dec 2016], [VOL Nov 2016])
) pvtvol
WHERE CustName='Bob'

这给出了预期的结果

CustName | OrderedProduct | VOL Dec 2016 | VOL Nov 2016
Bob      | Canned Tuna    |           16 |            4

我想要的是还要转移保证金，以便将 src 查询更改为：

    SELECT CustName, OrderedProduct, QtyOrdered,
    ((UnitPrice-UnitCost)*QtyOrdered) AS Margin,
    'VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthVol
    'MAR ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthMar
    FROM [ORDERS]

我尝试了以下修改后的查询添加第二个数据透视

SELECT CustName, OrderedProduct, [VOL Dec 2016], [VOL Nov 2016], [MAR Dec 2016], [MAR Nov 2016] FROM
(
    SELECT CustName, OrderedProduct, QtyOrdered,
    ((UnitPrice-UnitCost)*QtyOrdered) AS Margin,
    'VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthVol
    'MAR ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthMar
    FROM [ORDERS]
) src
PIVOT
(
    SUM(QtyOrdered)
    FOR YearMonthVol IN ([VOL Dec 2016], [VOL Nov 2016])
) pvtvol
PIVOT
(
    SUM(Margin)
    FOR YearMonthMar IN ([MAR Dec 2016], [MAR Nov 2016])
) pvtmar
WHERE CustName='Bob'

EXPECT 要查看的表格是

CustName | OrderedProduct | VOL Dec 2016 | VOL Nov 2016 | MAR Dec 2016 | MAR Nov 2016
Bob      | Canned Tuna    |           16 |            4 |           48 |            4

我实际的表格是

CustName | OrderedProduct | VOL Dec 2016 | VOL Nov 2016 | MAR Dec 2016 | MAR Nov 2016
Bob      | Canned Tuna    |           16 |         NULL |           48 |         NULL
Bob      | Canned Tuna    |         NULL |            4 |         NULL |            4

因此，似乎没有应用保证金上的支点，我为每个订单获得了一行。是我试图通过我尝试使用的方法实现可能实现的，或者我会更好地使用VOL和MAR的交叉表查询或者枢轴和交叉表的混合（VOL的枢轴， MAR的交叉表？）

任何建议都非常感谢。

Answer 1

您可以稍微简化一下您的查询

示例

Select *
 From (
        Select A.CustName 
              ,A.OrderedProduct
              ,B.*
         From  ORDERS A
         Cross Apply ( values ('VOL '+left(DateName(Month,OrderDate),3)+' '+DateName(Year,OrderDate), QtyOrdered )
                             ,('MAR '+left(DateName(Month,OrderDate),3)+' '+DateName(Year,OrderDate),((UnitPrice-UnitCost)*QtyOrdered) )
                     ) B (Item,Value)
         Where CustName='Bob'
      ) A
 Pivot (sum([Value]) For [Item] in ([VOL Dec 2016], [VOL Nov 2016],[MAR Dec 2016], [MAR Nov 2016]) ) p

<强>返回

Answer 2

您需要移动和转动

SELECT CustName, OrderedProduct, [VOL Dec 2016], [VOL Nov 2016], [MAR Dec 2016], [MAR Nov 2016] FROM
(
    SELECT CustName, OrderedProduct,YearMonth,val
    FROM [ORDERS]
    cross apply (values ('VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120),QtyOrdered),('MAR ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120),((UnitPrice-UnitCost)*QtyOrdered))) tc
    (YearMonth,val)
    WHERE CustName='Bob'
) src
PIVOT
(
    SUM(val)
    FOR YearMonth IN ([VOL Dec 2016], [VOL Nov 2016],[MAR Dec 2016], [MAR Nov 2016])
) pvtvol

• 的 Demo

结果中的重复记录是由于在透视源查询中存在QtyOrdered和Margin。必须单独调整QtyOrdered和Margin以避免重复。它必须像这样完成

;with QtyOrdered as
(
SELECT CustName, OrderedProduct, [VOL Dec 2016], [VOL Nov 2016]FROM
(
    SELECT CustName, OrderedProduct, QtyOrdered,
    'VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthVol
    FROM [ORDERS]
    WHERE CustName='Bob'
) src
PIVOT
(
    SUM(QtyOrdered)
    FOR YearMonthVol IN ([VOL Dec 2016], [VOL Nov 2016])
) pvtvol
),Margin as 
(
    select * from (
SELECT CustName, OrderedProduct, ((UnitPrice-UnitCost)*QtyOrdered) AS Margin,
    'MAR ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthMar
    FROM [ORDERS]
    ) a
PIVOT
(
    SUM(Margin)
    FOR YearMonthMar IN ([MAR Dec 2016], [MAR Nov 2016])
) pvtmar
WHERE CustName='Bob'    
)
select * from QtyOrdered q
join Margin m on q.CustName = m.CustName 
and q.OrderedProduct = m.OrderedProduct 

Answer 3

SELECT CustName, OrderedProduct, [VOL Dec 2016], [VOL Nov 2016], [MAR Dec 2016], [MAR Nov 2016] FROM
(
    SELECT CustName, OrderedProduct, QtyOrdered,
    ((UnitPrice-UnitCost)*QtyOrdered) AS Margin,
    'VOL ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthVol
    'MAR ' + CONVERT(CHAR(4), OrderDate, 100) + CONVERT(CHAR(4), OrderDate, 120) AS YearMonthMar
    FROM [ORDERS]
) src
PIVOT
(
    SUM(QtyOrdered)  qty, SUM(Margin) margin
    FOR YearMonthVol IN ([VOL Dec 2016], [VOL Nov 2016])
) pvtvol

WHERE CustName='Bob'