SQL Server中的运行总计无法正确计算

Question

我正在尝试实现以下运行总计，如下所示。其中a2013是该月的查询数量，2013c是运行总数。

Month   a2013   2013C
---------------------    
Jan       1       1    
Feb       3       4
March     2       6

在类似的帖子中，它建议添加我已经完成的分区，但返回的值仍然不正确。

with years as (
select datename(month,p.enquirydate) as Month, count(*) as count, 
count(case when p.enquirydate between '2013-01-01' and '2013-12-31 
23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2013,
count(case when p.enquirydate between '2014-01-01' and '2014-12-31 
23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2014,
count(case when p.enquirydate between '2015-01-01' and '2015-12-31 
23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2015,
count(case when p.txtenquirydate between '2016-01-01' and '2016-12-31 
23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2016,
count(case when p.enquirydate between '2017-01-01' and '2017-12-31 
23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2017
from customers p 
group by datename(month,p.enquirydate), datepart(month,p.enquirydate) 
)
select y.Month,a2013, sum(a2013) over(partition by y.month) as '2013C'
from years y

Month   a2013 2013C
NULL     0     0
April    4     4
August   1     1
December 4     4
February 3     3
January  1     1
July     3     3
June     3     3
March    2     2
May      1     1
November 4     4
October  3     3
September 4    4

Answer 1

也许以下查询可以帮助您使用SUM（）和UnBounded Preceding子句在SQL Server上实现SQL Running Sum解决方案

我曾经用于库存老化问题

declare @t as table (Id int, Month varchar(100), a2013 int)
insert into @t values 
(1,NULL     ,0),
(4,'April',    4 ), 
(8,'August',   1  ),
(12,'December', 4  ),
(2,'February', 3    ), 
(1,'January',  1     ),
(7,'July',     3     ),
(6,'June',     3     ),
(3,'March',    2     ),
(5,'May',      1     ),
(11,'November', 4     ),
(10,'October',  3     ),
(9,'September', 4    )



select
*,
sum([a2013]) over (order by Id rows unbounded preceding) as [2013c]
from @t
order by Id

如您所见，最后一列显示的是

之前显示的行数

Answer 2

with years as (
select DATEPART(month,enquirydate) as MonthNumber, count(*) as count, 
count(case when p.enquirydate between '2013-01-01' and '2013-12-31 23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2013,
count(case when p.enquirydate between '2014-01-01' and '2014-12-31 23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2014,
count(case when p.enquirydate between '2015-01-01' and '2015-12-31 23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2015,
count(case when p.enquirydate between '2016-01-01' and '2016-12-31 23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2016,
count(case when p.enquirydate between '2017-01-01' and '2017-12-31 23:59:59.997' and EnrolmentYear ='2018' then datename(month,p.enquirydate) 
else null end) as a2017
from @customers p 
group by DATEPART(month,enquirydate)
)
select datename(month,dateadd(month, MonthNumber - 1, 0)), a2013, sum(a2013) over(Order By y.MonthNumber ROWS UNBOUNDED PRECEDING) as '2013C'
from years y