TSQL：从层次结构的顶部到底部的总和

我正在尝试做一些关于多层次结构的运行总计

样本表。我添加了空格来强调分组

+----------------+----------------+------------+
|    Manager     |    Employee    |    Cost    |
+----------------+----------------+------------+
| Donald Torres  | Pamela Jordan  | 7,250.78   |
|                |                |            |
| Pamela Jordan  | Harry Rivera   | 36,721.39  |
| Pamela Jordan  | Ryan Mcdonald  | 5,593.89   |
|                |                |            |
| Harry Rivera   | Bruce Mccoy    | 3,538.19   |
|                |                |            |
| Ryan Mcdonald  | Shirley Ortiz  | 137.20     |
| Ryan Mcdonald  | Roger Gardner  | 9,140.34   |
| Ryan Mcdonald  | Paula Richards | 135,109.59 |
|                |                |            |
| Paula Richards | Andrew Boyd    | 106,676.77 |
+----------------+----------------+------------+

期望输出

+----------------+----------------+------------+
|    Manager     |    Employee    |    Cost    |
+----------------+----------------+------------+
| Donald Torres  | Pamela Jordan  | 304,168.15 |
| Pamela Jordan  | Harry Rivera   | 40,259.58  |
| Pamela Jordan  | Ryan Mcdonald  | 256,657.79 |
| Harry Rivera   | Bruce Mccoy    | 3,538.19   |
| Ryan Mcdonald  | Shirley Ortiz  | 137.20     |
| Ryan Mcdonald  | Roger Gardner  | 9,140.34   |
| Ryan Mcdonald  | Paula Richards | 241,786.36 |
| Paula Richards | Andrew Boyd    | 106,676.77 |
+----------------+----------------+------------+

为给出不正确的期望输出而道歉。我已经更新了表格

基本上，这是基于初始表的层次结构的外观：

Donald Torres                   
    Pamela Jordan               7,250.78
        Harry Rivera            36,721.39
            Bruce Mccoy         3,538.19
        Ryan Mcdonald           5,593.89
            Shirley Ortiz       137.20
            Roger Gardner       9,140.34
            Paula Richards      135,109.59
                Andrew Boyd     106,676.77

关于期望的产出，Paula Richards应该得到Paula Richards（她的成本）+Andrew Boyd的总和。Ryan Mcdonald将把Ryan Mcdonald（hir成本）的总额减至Andrew Boyd。哈里·里维拉只会是哈里·里维拉+布鲁斯·麦考伊。最后，帕梅拉·乔丹将把帕梅拉·乔丹（她的费用）的总额减至安德鲁·博伊德

我正在尝试用SQL 2008R2编写查询

更新：添加cte基表

With Tabl_1 (Manager, Employee, Cost) as (
    Select 'Donald Torres'  , 'Pamela Jordan'  , 7250.78   UNION ALL
    Select 'Pamela Jordan'  , 'Harry Rivera'   , 36721.39  UNION ALL
    Select 'Pamela Jordan'  , 'Ryan Mcdonald'  , 5593.89   UNION ALL
    Select 'Harry Rivera'   , 'Bruce Mccoy'    , 3538.19   UNION ALL
    Select 'Ryan Mcdonald'  , 'Shirley Ortiz'  , 137.20     UNION ALL
    Select 'Ryan Mcdonald'  , 'Roger Gardner'  , 9140.34   UNION ALL
    Select 'Ryan Mcdonald'  , 'Paula Richards' , 135109.59 UNION ALL
    Select 'Paula Richards' , 'Andrew Boyd'    , 106676.77
)

---

我认为这个问题符合你的需要

With Table_1 (Manager, Employee, Cost) as (
    Select 'Donald Torres'  , 'Pamela Jordan'  , 7250.78   UNION ALL
    Select 'Pamela Jordan'  , 'Harry Rivera'   , 36721.39  UNION ALL
    Select 'Pamela Jordan'  , 'Ryan Mcdonald'  , 5593.89   UNION ALL
    Select 'Harry Rivera'   , 'Bruce Mccoy'    , 3538.19   UNION ALL
    Select 'Ryan Mcdonald'  , 'Shirley Ortiz'  , 137.20     UNION ALL
    Select 'Ryan Mcdonald'  , 'Roger Gardner'  , 9140.34   UNION ALL
    Select 'Ryan Mcdonald'  , 'Paula Richards' , 135109.59 UNION ALL
    Select 'Paula Richards' , 'Andrew Boyd'    , 106676.77
),
CTE as
    (
        select Manager as Manager, Employee as Employee, 1 as Level, Cost as Cost
        from Table_1
        Union All
        select C.Manager, T.Employee, C.Level + 1, T.Cost
        from CTE as C inner join Table_1 as T on C.Employee = T.Manager
    )

select Manager, 
       Employee, 
       T_Cost = Cost + isnull((select SUM(Cost) 
                               from CTE C2 
                               where C2.Manager = C1.Employee)
                             ,0) 
from CTE C1 
where Level = 1

---

期望的输出非常混乱。为什么所有其他组都有一个非常大的累积值，而哈利/布鲁斯系列却没有？为什么瑞安·麦克唐纳/罗杰·加德纳没有包括雪莉·奥尔蒂斯的值？您如何确定每位经理的员工顺序？嗨@BaconBits。确定每位经理的员工顺序也是我的问题。不确定使用递归queryHi@Polux2是否可以实现这一点。抱歉，它没有得到所需的输出。@ggarcia您是对的，我已经更正了我的查询，抱歉。