三种实现方法实现数据表中遍历寻找子节点
示例问题如下：
表结构:
Id ParentId
1 0
2 1
3 2
......
针对该表结构解释如下:
1的父节点为0,
2的父节点为1,
3的父节点为2
......
以此类推,要求给定一个父节点的值,比如1,
用SQL语句查询的到该父结点下的所有子节点
下面的Sql是在sql server下调试通过的,如果是oracle,则有Connect By可以实现.
建立测试表:
Drop table DbTree
Create table DbTree
(
[Id] Int,
[Name] NVarChar(20),
[ParentId] Int
)
插入测试数据:
Insert Into DbTree ([Id],[ParentId]) Values (1,0)
Insert Into DbTree ([Id],[ParentId]) Values (2,1)
Insert Into DbTree ([Id],[ParentId]) Values (3,1)
Insert Into DbTree ([Id],[ParentId]) Values (4,3)
Insert Into DbTree ([Id],[ParentId]) Values (5,4)
Insert Into DbTree ([Id],[ParentId]) Values (6,7)
Insert Into DbTree ([Id],[ParentId]) Values (8,5)
实现方法一:
代码如下:
Declare @Id Int
Set @Id = 1 ---在次修改父节点
Select * Into #Temp From DbTree Where ParentId In (@Id)
Select * Into #AllRow From DbTree Where ParentId In (@Id) --1,2
While Exists(Select * From #Temp)
Begin
Select * Into #Temp2 From #Temp
Truncate table #Temp
Insert Into #Temp Select * From DbTree Where ParentId In (Select Id From #Temp2)
Insert Into #AllRow Select * From #Temp
Drop table #Temp2
End
Select * From #AllRow Order By Id
Drop table #Temp
Drop table #AllRow
实现方法二:代码如下:
Create table #AllRow
(
Id Int,
ParentId Int
)
Declare @Id Int
Set @Id = 1 ---在次修改父节点
Delete #AllRow
--顶层自身
Insert Into #AllRow (Id,ParentId) Select @Id, @Id
While @@RowCount > 0
Begin
Insert Into #AllRow (Id,ParentId)
Select B.Id,A.Id
From #AllRow A,DbTree B
Where A.Id = B.ParentId And
Not Exists (Select Id From #AllRow Where Id = B.Id And ParentId = A.Id)
End
Delete From #AllRow Where Id = @Id
Select * From #AllRow Order By Id
Drop table #AllRow
实现方法三:
代码如下:
在sql server2005中其实提供了CTE[公共表表达式]来实现递归:
关于CTE的使用请查MSDN
Declare @Id Int
Set @Id = 3; ---在次修改父节点
With RootNodeCTE(Id,ParentId)
As
(
Select Id,ParentId From DbTree Where ParentId In (@Id)
Union All
Select DbTree.Id,DbTree.ParentId From RootNodeCTE
Inner Join DbTree
On RootNodeCTE.Id = DbTree.ParentId
)
Select * From RootNodeCTE