SQL在定义的时间范围内将累积值分解为单个值

如何解决SQL在定义的时间范围内将累积值分解为单个值

我有一张图，如下图1所示，我试图将每个日期和每个部门的累计收入分解为单个利润。

但是，我使用的方法非常愚蠢，如果日期范围过大（即将收入从2020年1月1日拆分为2020年7月31日），则无法应用。此外，结果也不是我所期望的，因为单个利润是按列而不是按行显示的。

即使日期范围很大，有没有一种更好的方法可以使我得到如图2所示的预期结果？

我当前的解决方案：

declare @t1 varchar(10)
declare @t2 varchar(10)
declare @t3 varchar(10)
set @t1 = '20200701'
set @t2 = '20200702'
set @t3 = '20200703'


select M_Date,Department,[T1_revenue] as [Profit_t1],([T2_revenue] - [T1_revenue]) as [Profit_t2],([T3_revenue] - [T2_revenue]) as [Profit_t3]
from(
    select M_Date,Cumulated_Revenues as [T1_revenue]
    from dmf_dbo.profitTable
    WHERE M_Date = @t1
    left join (
        select M_Date,Cumulated_Revenues as [T2_revenue]
        from dmf_dbo.profitTable
        WHERE M_Date = @t2
    ) AS p_t2
    on M_Date = p_t2.M_Date and  Department = p_t2.Department
    left join (
        select M_Date,Cumulated_Revenues as [T3_revenue]
        from dmf_dbo.profitTable
        WHERE M_Date = @t3
    ) AS p_t3
    on p_t2.M_Date = p_t3.M_Date and  p_t2.Department = p_t3.Department
)

图片1.原始数据：

图片2.预期结果：

解决方法

我想我明白了。您只希望lag()使用默认值：

select t.*,(cumulative_revenue -
        lag(cumulative_revenue,1,0) over (partition by department order by m_date)
       ) as profit
from t;