Sql server 基于汇总类型的树状结构中的总分

我试图计算树结构中每个节点的总计，计算总计取决于汇总类型

我有如下结构

• 如果节点类型为当前列表=查看当前节点，则无需在树中向下搜索
• 如果节点类型为子节点，则忽略当前节点并向下
• 如果节点类型为“当前”且子节点为“当前”，则包含当前节点 然后下去

假设我的起点是B1：那么总分是=（20+40+110+60+90）

假设我的起点C2：分数是110

我设法把所有的孩子都弄到手了，但到了集合的时候，我被卡住了，请帮帮我

示例数据脚本

+--------+----------+------+------------+-------------------+------------------+
| Teamid | ParentId | Name | ParentName |       Type        | Key Result Value |
+--------+----------+------+------------+-------------------+------------------+
|      1 | null     | A1   | Null       | Current and Child |               10 |
|      2 | 1        | B1   | A1         | Current and Child |               20 |
|      3 | 1        | B2   | A1         | Current and Child |               30 |
|      4 | 2        | C1   | B1         | Current List      |               40 |
|      5 | 2        | C2   | B1         | Child List        |               50 |
|      6 | 2        | C3   | B1         | Current and Child |               60 |
|      7 | 4        | D1   | C1         | Current and Child |               70 |
|      8 | 5        | D2   | C2         | Child List        |               80 |
|      9 | 6        | D3   | C3         | Current and Child |               90 |
|     10 | 7        | E1   | D1         | Current List      |              100 |
|     11 | 8        | E2   | D2         | Current and Child |              110 |
+--------+----------+------+------------+-------------------+------------------+

GroupName  GroupSum  GroupSumFormula
---------  --------  ---------------------
A1         360       10+20+30+40+60+90+110
B1         320       20+40+60+90+110
B2         30        30
C1         40        40
C2         110       110
C3         150       60+90
D1         170       100+70
D2         110       110
D3         90        90
E1         100       100
E2         110       110

树表脚本：

CREATE TABLE TreeTable(
   Teamid           INTEGER  NOT NULL PRIMARY KEY 
  ,ParentId         INTEGER 
  ,Name             VARCHAR(2) NOT NULL
  ,ParentName       VARCHAR(4)
  ,Type             VARCHAR(17) NOT NULL
  ,Key_Result_Value INTEGER  NOT NULL
);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (1,NULL,'A1',NULL,'Current and Child',10);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (2,1,'B1','A1','Current and Child',20);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (3,1,'B2','A1','Current and Child',30);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (4,2,'C1','B1','Current List',40);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (5,2,'C2','B1','Child List',50);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (6,2,'C3','B1','Current and Child',60);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (7,4,'D1','C1','Current and Child',70);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (8,5,'D2','C2','Child List',80);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (9,6,'D3','C2','Current and Child',90);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (10,7,'E1','D1','Current List',100);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (11,8,'E2','E2','Current and Child',110);

WITH
CTE_Groups
AS
(
    SELECT
        TreeTable.teamid AS StartID
        ,TreeTable.Name AS StartName
        ,TreeTable.teamid
        ,TreeTable.ParentID
        ,TreeTable.Name
        ,1 AS Lvl
        ,TreeTable.Type
        ,TreeTable.Key_Result_Value
    FROM TreeTable
    
    UNION ALL

    SELECT
        CTE_Groups.StartID
        ,CTE_Groups.StartName
        ,TreeTable.teamid
        ,TreeTable.ParentID
        ,TreeTable.Name
        ,CTE_Groups.Lvl + 1 AS Lvl
        ,TreeTable.Type
        ,TreeTable.Key_Result_Value

    FROM
        TreeTable
        INNER JOIN CTE_Groups ON CTE_Groups.teamid = TreeTable.ParentID
)
select * from CTE_Groups
order by startid,lvl

我的尝试：

CREATE TABLE TreeTable(
   Teamid           INTEGER  NOT NULL PRIMARY KEY 
  ,ParentId         INTEGER 
  ,Name             VARCHAR(2) NOT NULL
  ,ParentName       VARCHAR(4)
  ,Type             VARCHAR(17) NOT NULL
  ,Key_Result_Value INTEGER  NOT NULL
);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (1,NULL,'A1',NULL,'Current and Child',10);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (2,1,'B1','A1','Current and Child',20);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (3,1,'B2','A1','Current and Child',30);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (4,2,'C1','B1','Current List',40);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (5,2,'C2','B1','Child List',50);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (6,2,'C3','B1','Current and Child',60);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (7,4,'D1','C1','Current and Child',70);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (8,5,'D2','C2','Child List',80);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (9,6,'D3','C2','Current and Child',90);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (10,7,'E1','D1','Current List',100);
INSERT INTO TreeTable(Teamid,ParentId,Name,ParentName,Type,Key_Result_Value) VALUES (11,8,'E2','E2','Current and Child',110);

WITH
CTE_Groups
AS
(
    SELECT
        TreeTable.teamid AS StartID
        ,TreeTable.Name AS StartName
        ,TreeTable.teamid
        ,TreeTable.ParentID
        ,TreeTable.Name
        ,1 AS Lvl
        ,TreeTable.Type
        ,TreeTable.Key_Result_Value
    FROM TreeTable
    
    UNION ALL

    SELECT
        CTE_Groups.StartID
        ,CTE_Groups.StartName
        ,TreeTable.teamid
        ,TreeTable.ParentID
        ,TreeTable.Name
        ,CTE_Groups.Lvl + 1 AS Lvl
        ,TreeTable.Type
        ,TreeTable.Key_Result_Value

    FROM
        TreeTable
        INNER JOIN CTE_Groups ON CTE_Groups.teamid = TreeTable.ParentID
)
select * from CTE_Groups
order by startid,lvl

• 我试图获取节点的所有子节点的屏幕截图，但后来我被困在如何根据类型排除节点上，因为我需要了解完整的路径：

请帮忙 预期结果：

+--------+-----------------------+
| TeamID |         Score         |
+--------+-----------------------+
| E1     | 100                   |
| D1     | 70 + 100              |
| C1     | 40                    |
| E2     | 110                   |
| D2     | 110                   |
| C2     | 110                   |
| D3     | 90                    |
| C3     | 60+90                 |
| B1     | 20+40+110+60+90       |
| B2     | 30                    |
| A1     | 10+30+20+40+110+60+90 |
+--------+-----------------------+

---

使用递归列出节点子节点是一个良好的开端。使用

case

表达式，您还可以跟踪组类型以及何时添加或忽略值（忽略=使用

0

而不是实际节点值）

解决方案

with rcte as
(
  select t.Name as GroupName,
         t.Type as GroupType,
         t.Name,
         t.Type,
         case t.Type
           when 'Child list' then 0 -- if the current node must be ignored, then the current value changes to 0
           else t.Key_Result_Value
         end as Value
  from TreeTable t
union all
  select r.GroupName,
         r.GroupType,
         t.Name,
         case r.Type
           when 'Current list' then 'Current list' -- if the parent was the end of the sum, then it stays the end of the sum
           else t.Type
         end,
         case
           when r.Type = 'Current list' then 0 -- if the parent was the end of the sum, then the current value changes to 0
           when t.Type = 'Child list' then 0 -- if the current node must be ignored, then the current value changes to 0
           else t.Key_Result_Value
         end
  from rcte r
  join TreeTable t
    on t.ParentName = r.Name
)
select r.GroupName,
       sum(r.Value) as GroupSum,
       string_agg(case when r.Value <> 0 then convert(nvarchar(3), r.Value) end, '+') as GroupSumFormula
from rcte r
group by r.GroupName
order by r.GroupName;

---

查看实际情况。

您是只想对D1

170

的总数进行汇总，还是需要知道单个总数

70+100

示例数据脚本中存在键入错误。

E2

的父项是

D2

而不是它本身……示例数据脚本中还有另一个输入错误。

D3

的父项是

C3

，而不是

C2

@Charlieface170@Sander我修好了it@sara你的树中有一个循环，在那里抱歉使用了带有打字错误的脚本。它的工作原理就像一个charm@Sander如果情况恰好是多亲，那么当起点为B1时，为什么B1会寻找其父代并获得重复分数？如果B1有多亲，那么这些父代中的每一个都会包含B1的子树值。这样的需求是一个很大的变化，当前的解决方案无法正确处理。请发布一个新问题，其中包含多个家长的数据样本和预期结果。一个新问题会比一个已经回答的问题得到更多的关注。@Sander你能研究一下这个问题吗