将特定列转换为 SQL 中的行

Question

此屏幕截图包含我的结果，我想通过将AMOUNT_BILLED_JANUARY 列转换为如下行来进一步转置此查询：

M_ACCOUNTCONTRACT	Bill_Date	Previous1	Previous2	Previous3
16959	2019-05-13	10670.75	8580.66	6927.59
16959	2019-04-09	8580.66	6927.59	8181.64
16959	2019-03-09	6927.59	8181.64	8751.67

Answer 1

将您当前的查询移动到公用表表达式 (cte) 中，但将 row_number() 函数的返回值保留为整数。您可以使用整数值来获取前面的行。

样本数据

复制当前结果的开头，但添加了一列。

create table currentResult
(
  M_AccountContract int,
  Bill_Date date,
  Amount_Billed_January decimal(7,2),
  Row_Num nvarchar(15),
  Num int -- include the row_number() function as integer value
);

insert into currentResult (M_AccountContract, Bill_Date, Amount_Billed_January, Row_Num, Num) values
(16959, '2019-05-13', 10670.75, 'previous1', 1),
(16959, '2019-04-09',  8580.66, 'previous2', 2),
(16959, '2019-03-12',  6927.59, 'previous3', 3),
(16959, '2019-02-08',  8181.64, 'previous4', 4),
(16959, '2019-01-11',  8751.67, 'previous5', 5);
-- got lazy here, please post text instead of images next time ;)

解决方案

使用您当前的结果并将其与自身连接，以查找必须通过一些简单的数学 (c2.Num >= c1.Num and c2.Num < c1.Num + 3) 进行旋转的行组，并在每个组中生成一个新的行号 (PrevNum)。然后使用新的PrevNum 列旋转每个组。

with cte as
(
  -- current query with row_number() integer value goes here
  select *
  from currentResult
),
cte_prev3 as
(
  select c1.M_AccountContract,
         c1.Bill_Date,
         row_number() over(partition by c1.M_AccountContract, c1.Bill_Date order by c2.Num) as PrevNum,
         c2.Amount_Billed_January
  from cte c1
  left join cte c2
    on  c2.M_AccountContract = c1.M_AccountContract
    and c2.Num >= c1.Num
    and c2.Num <  c1.Num + 3
)
select piv.M_AccountContract,
       piv.Bill_Date,
       piv.[1] as Previous1,
       piv.[2] as Previous2,
       piv.[3] as Previous3
from cte_prev3 cp3
pivot (max(cp3.Amount_Billed_January) for cp3.PrevNum in ([1], [2], [3])) piv
order by piv.Bill_Date desc;

结果

null 值是有限样本数据的结果。

M_AccountContract  Bill_Date   Previous1  Previous2  Previous3
-----------------  ----------  ---------  ---------  ---------
16959              2019-05-13   10670.75    8580.66    6927.59
16959              2019-04-09    8580.66    6927.59    8181.64
16959              2019-03-12    6927.59    8181.64    8751.67
16959              2019-02-08    8181.64    8751.67       null
16959              2019-01-11    8751.67       null       null

Fiddle 以查看实际情况以及中间的cte_prev3 结果。