Question

这是我的基础表：

Region year carsSold bicyclesSold
RegX   1999       50          100
RegX   2000      150           30
RegY   1998       60           40

我希望得到的结果是：

Region Year          1998 1999 2000
RegX   carsSold         0   50  150
       bicyclesSold     0  100   30  
RegY   carsSold        60    0    0
       bicyclesSold    40    0    0

有没有办法在sql server中执行此操作？有没有办法在Excel中执行此操作？这种数据表示的最佳解决方案是什么。你能建议任何适合的软件吗？

我知道在sql server中透视表，但是不能写任何能让结果接近我想要的东西

Answer 1

select P.Region,
       P.Item,
       isnull(P.[1998], 0) as [1998],
       isnull(P.[1999], 0) as [1999],
       isnull(P.[2000], 0) as [2000]
from YourTable
unpivot (ItemsSold for Item in (carsSold, bicyclesSold)) as U
pivot (sum(ItemsSold) for year in ([1998], [1999], [2000])) as P
order by P.Region, P.Item

SQL Fiddle

Answer 2

作为替代方案，仅限ANSI-92兼容代码。哎呀，你甚至可以在SQLite中运行它，除了关于对列名称进行方括号的部分。

当你说exactly what I want

时，我带你字面

select
  case when veh.iscar=1 then t.region else '' end Region,
  veh.type [Year],
  max(case when t.year = 1998
      then carsSold*iscar + bicyclesSold*(1-iscar) else 0 end) [1998],
  max(case when t.year = 1999
      then carsSold*iscar + bicyclesSold*(1-iscar) else 0 end) [1999],
  max(case when t.year = 2000
      then carsSold*iscar + bicyclesSold*(1-iscar) else 0 end) [2000]
from YourTable t
cross join (
  select 'carsSold',1 union all
  select 'bicyclesSold',0) veh(type,iscar)
group by t.region, veh.type, veh.iscar
order by t.region, type desc;

>>
Region  Year         1998   1999    2000
RegX    carsSold        0   50      150
        bicyclesSold    0   100     30
RegY    carsSold        60  0       0
        bicyclesSold    40  0       0

透视表并获取多个参数的聚合

2 个答案: