SQL查询(可能与pivot/unpivot解决？)

Question

我不确定如何最好地解决这个问题 - 我认为应该使用pivot/unpivot,但不确定如何使其工作(因为旋转列是非数字的)

我有下表(基于我无法修改的查询):

CREATE TABLE #data 
(donor_id NVARCHAR(50)
,last_gift DATETIME
,[2005] NVARCHAR(50)
,numgifts05 INT
,value_05 MONEY
,[2006] NVARCHAR(50)
,numgifts06 INT
,value_06 MONEY
,[2007] NVARCHAR(50)
,numgifts07 INT
,value_07 MONEY
,[2008] NVARCHAR(50)
,numgifts08 INT
,value_08 MONEY
,[2009] NVARCHAR(50)
,numgifts09 INT
,value_09 MONEY
,[2010] NVARCHAR(50)
,numgifts10 INT
,value_10 MONEY
)


INSERT INTO #data VALUES  (001,'2000-03-23 00:00:00.000','lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (002,'2006-08-01 00:00:00.000','reactivated donor',1,25.00,'2yrs consecutive',2,47.20,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (003,'2010-02-12 00:00:00.000','non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'new donor',2,112.50,'2yrs consecutive',1,116.08)
INSERT INTO #data VALUES  (004,'2010-01-04 00:00:00.000','non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'new donor',1,12.00)
INSERT INTO #data VALUES  (005,'2009-12-09 00:00:00.000','non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'new donor',1,18.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (006,'2005-01-10 00:00:00.000','new donor',1,20.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (007,'2009-01-07 00:00:00.000','non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'non-donor',0,0.00,'new donor',1,25.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (008,'1997-11-05 00:00:00.000','lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (009,'1999-06-18 00:00:00.000','lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00,'lapsed donor',0,0.00)
INSERT INTO #data VALUES  (010,'2010-03-09 00:00:00.000','3/4yrs consecutive',12,120.00,'3/4yrs consecutive',12,120.00,'5yrs+ consecutive',13,142.40,'5yrs+ consecutive',13,170.00,'5yrs+ consecutive',13,160.00,'5yrs+ consecutive',3,30.00)

我需要生成的是一个看起来像这样的表:

Category          |  2005   |  2006  |  2007  | 2008  |  2009  |  2010

Lapsed donor      | count(lapsed for 2005) |
New donor         | count(new donor for 2005) |
2yrs consecutive  | count(2yrs consecutive for 2005) |
etc...

然而,我遇到的问题是,如果某人在2005年失去了捐助者,他们在2006 - 2010年不一定属于那一类,那么每个人有可能每年出现在不同的一行吗？

我知道所有的类别选项都是固定的,因此不需要动态旋转或类似的东西,以防它有帮助:)

主要目标是能够一目了然地看到每一年的每个类别的数量(所以我将所有年份相互比较) - 任何帮助都非常赞赏!!

Answer 1

这是你需要的吗？

SELECT Category,[2005],[2006],[2007],[2008],[2009],[2010]  
FROM 
(SELECT [2005],[2006],[2007],[2008],[2009],[2010],donor_id FROM #DATA) P
UNPIVOT  (Category FOR [YEAR] IN 
      ([2005],[2006],[2007],[2008],[2009],[2010])
)AS unpvt
PIVOT
(
COUNT (donor_id)
FOR [YEAR] IN
( [2005],[2006],[2007],[2008],[2009],[2010])
) AS pvt
ORDER BY Category

返回

Category                       2005        2006        2007        2008        2009        2010
------------------------------ ----------- ----------- ----------- ----------- ----------- -----------
2yrs consecutive               0           1           0           0           0           1
3/4yrs consecutive             1           1           0           0           0           0
5yrs+ consecutive              0           0           1           1           1           1
lapsed donor                   3           4           5           5           5           7
new donor                      1           0           0           0           3           1
non-donor                      4           4           4           4           1           0
reactivated donor              1           0           0           0           0           0

澄清之后

With base AS
(
SELECT 2005 as [Year], [2005] as Category, numgifts05 as numgifts,  value_05  as value
from #data
union all
SELECT  2006 as [Year], [2006] as Category, numgifts06 as numgifts,  value_06  as value
from #data
)
select 
    Category
    ,COUNT(case when [Year]=2005 then 1 end) as [2005]
    ,SUM(case when [Year]=2005 then numgifts end) as [numgifts05]
    ,SUM(case when [Year]=2005 then value end) as [value_05]
    ,COUNT(case when [Year]=2006 then 1 end) as [2006]
    ,SUM(case when [Year]=2006 then numgifts end) as [numgifts06]
    ,SUM(case when [Year]=2006 then value end) as [value_06]
from base
group by Category
order by Category