这是我的CustomerDetails表。
CustomerID CustCodeID
25 1
65 8
35 2
112 8
45 2
975 8
364 1
48 8
69 1
97 8
33 1
11 8
93 2
10 8
21 1
65 8
74 2
53 8
这是我的Fact_SalesMetrics表。
Date Sales # CustomerID
2015-03-23 00:00:00.000 42895 25
2015-03-13 00:00:00.000 53920 53
2015-03-23 00:00:00.000 44895 65
2015-03-13 00:00:00.000 43920 35
2015-03-23 00:00:00.000 48895 112
2015-03-13 00:00:00.000 47920 45
2015-03-23 00:00:00.000 46895 975
2015-03-13 00:00:00.000 45920 48
2015-03-23 00:00:00.000 40895 69
2015-03-13 00:00:00.000 40920 11
2015-03-23 00:00:00.000 41895 33
2015-03-13 00:00:00.000 49920 21
......
我希望输出如下：
CustCodeID March 2015
1 4
2 2
8 7
这意味着拥有codeID＆＃39; 1＆＃39; 3月份有4个订单，2个订单有2个订单。
为了实现这一目标，我在下面查询并确保其正常运行：
select CustCodeID,sum(March) as 'March 2015' from (
select bb.CustCodeID, aa.March from (
(SELECT count(distinct([Sales #])) as 'March', customerid
FROM [SalesData].[dbo].[Fact_SalesMetrics] a
where date >= '2015-03-01 00:00:00.000' and date <= '2015-03-31 00:00:00.000'
and customerid in (select customerid from CustomerDetails)
group by customerid ) as aa inner join (select customerid,CustCodeID from CustomerDetails ) as bb on aa.customerid=bb.customerid
)
) as dd group by CustCodeID
现在我想计算过去三个月的发票数，如下所示：
CustCodeID March 2015 February 2015 January 2015
1 4 ? ?
2 2 ? ?
8 7 ? ?
任何人都可以帮我实现这个目标吗？
答案 0 :(得分：0)
由于您拥有DateTime类型的列，因此需要将其转换为Month Year格式。在Sql Server中使用DateName来提取月份名和年份。然后，您应该使用新的日期格式和CustomerID找到CustCodeID和GROUP BY的计数。您应该使用此查询作为要转动的表的源查询。
如果预先知道月份的值，则可以通过对列名称进行硬编码来使用静态数据透视
SELECT CustCodeID,[March 2015],[February 2015],[January 2015]
FROM
(
SELECT CD.CustCodeID,COUNT(CD.CustomerID) COUNTOfCustomerID,
DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE]) MONTHS
FROM CustomerDetails CD
JOIN Fact_SalesMetrics FS ON CD.CustomerID=FS.CustomerID
GROUP BY CD.CustCodeID,DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE])
)TAB
PIVOT
(
MIN(COUNTOfCustomerID)
FOR MONTHS IN([March 2015],[February 2015],[January 2015])
)P
如果事先不知道列数，您可以选择动态数据透视。
为此，第一步是在从行进行透视后显示列名。在以下查询中，它将选择过去3个月的列。
DECLARE @cols NVARCHAR (MAX)
SELECT @cols = STUFF((SELECT ',' + QUOTENAME(MONTHS)
FROM
(
SELECT TOP 3 DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE]) MONTHS
from Fact_SalesMetrics
GROUP BY '01 ' + DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE]),
DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE])
ORDER BY CAST('01 ' + DATENAME(MONTH,[DATE])+' ' + DATENAME(YEAR,[DATE]) AS DATE) DESC
) c
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
现在使用Dynamic sql
执行数据透视查询DECLARE @query NVARCHAR(MAX)
SET @query = '
SELECT * FROM
(
SELECT CD.CustCodeID,COUNT(CD.CustomerID) COUNTOfCustomerID,
DATENAME(MONTH,[DATE])+'' '' + DATENAME(YEAR,[DATE]) MONTHS
FROM CustomerDetails CD
JOIN Fact_SalesMetrics FS ON CD.CustomerID=FS.CustomerID
GROUP BY CD.CustCodeID,DATENAME(MONTH,[DATE])+'' '' + DATENAME(YEAR,[DATE])
) x
PIVOT
(
-- Specify the values to hold in pivoted column
MIN([COUNTOfCustomerID])
-- Get the column names from variable
FOR [MONTHS] IN('+@cols+')
) p
ORDER BY CustCodeID;'
EXEC SP_EXECUTESQL @query