zoukankan      html  css  js  c++  java
  • Post Order traverse binary tree using non-recursive way

    Process analysis

        

    1. Stack = 5,  
    2. Push 3, Stack = 5, 3.    Pre = 5
    3. Current = 3, Pre = 5, Push 2 to the stack, stack = 5, 3, 2, Pre = 3
    4. Current = 2, Pre = 3, pop 2, PRINT 2, Stack = 5, 3.  Pre = 2
    5. Current = 3, push 4, Stack = 5, 3 , 4.  Pre = 3
    6. Current = 4,  4 is leaf node, pop 4.  PRINT 4, Pre = 4, Stack = 5, 3
    7. Current = 3, Pre = 4, 4 is right child of 3, it means it goes up from right side, all visited.  Pop one more element from stack, PRINT 3.  Pre = 3.  Stack = 5
    8. Current = 5, Pre = 3, 5 has right child, push 7, Stack = 5, 7, Pre = 5
    9. Current 7, which is right child of 5, meaning it's first time to visit 7, 7 has left child, push 6 to the stack.  Stack = 5, 7, 6    Pre = 7
    10. Current = 6, 6 is 7's left child, but 6 is leaf node, pop 6.  Print 6, Pre = 6. Stack = 5, 7
    11. Current = 7, 7 has right child?  yes, 8. Push 8 to the stack, Stack = 5,7,8.  Pre = 7
    12. Current = 8, 8 has no left child, has right child 10, push 10 to the stack.  Stack = 5, 7, 8, 10.  Pre = 8
    13. Current = 10, 10 is leaf node, pop 10, Print 10, Stack = 5,7,8, Pre = 10
    14. Current = 8, pop 8, Print 8, Stack = 5, 7, Pre = 8
    15. Current = 7, pop 7, Print 7, Stack = 5, Pre = 7
    16. Current = 5, pop 5, Print 5,  stack = empty
     public void PostOrderNonRecursion1StackV1()
            {
                Stack<TreeNode<T>> stack = new Stack<TreeNode<T>>();
    
                if (root == null)
                {
                    return;
                }
    
                TreeNode<T> current = this.root;
                TreeNode<T> pre = null;
    
                stack.Push(this.root);
    
                while (stack.Count != 0)
                {
                    current = stack.Peek();
    
                    // top down
                    if (pre == null || pre.leftChild == current || pre.rightChild == current)
                    {
                        if (current.leftChild != null)
                        {
                            stack.Push(current.leftChild);
                        }
                        else if (current.rightChild != null)
                        {
                            stack.Push(current.rightChild);
                        }
                        else
                        {
                            // leaf node, have to pop
                            Console.WriteLine("visit node {0}", stack.Pop().data);
                        }
                    }
                    // move up case from left node
                    else if (current.leftChild == pre)
                    {
                        if (current.rightChild != null)
                        {
                            stack.Push(current.rightChild);
                        }
                        else
                        {
                            Console.WriteLine("visit node {0}", stack.Pop().data);
                        }
                    }
                    // move up from right , which means all done, need pop up next one to figure out next action
                    else if (current.rightChild == pre)
                    {
                        Console.WriteLine("visit node {0}", stack.Pop().data);
                    }
    
                    // remember the previous node, not parent node
                    pre = current;
                }
            }

    The solution above uses one stack.  If you use two stacks, there are the following two ways.

    Solution 1 - do pre-order non-recursive firstly, but the sequence is root, right, left, not root, left and right.

    Then store the result in stack 2, print stack 2, you will get the final result.  Put code as below.

     public void PostOrderNonRecursion2StacksV1()
            {
                Stack<TreeNode<T>> stack1 = new Stack<TreeNode<T>>();
                Stack<TreeNode<T>> stack2 = new Stack<TreeNode<T>>();
    
                TreeNode<T> current = root;
    
                if (current == null)
                {
                    return;
                }
    
                while (current != null)
                {
                    stack2.Push(current);
                    Console.WriteLine(current.data);
                    stack1.Push(current);
                    current = current.rightChild;
                }
    
                while (stack1.Count != 0)
                {
                    current = stack1.Pop();
    
                    if (current.leftChild != null)
                    {
                        current = current.leftChild;
                        while (current != null)
                        {
                            stack2.Push(current);
                            stack1.Push(current);
                            current = current.rightChild;
                        }
                    }
                }
    
                while (stack2.Count != 0)
                {
                    Console.WriteLine("value = {0}", stack2.Pop().data);
                }
            }

    Solution 2:

    这个方法对于初学者很难想到,根节点先入栈,然后根节点出栈,存入栈2.  根节点的左孩子入栈,根节点的右孩子入栈。

    然后右孩子出栈,入栈2. 右孩子的左右节点入栈1。这样栈2的元素实际从底往上是主根,右孩子,递归右左。  反过来就是先左后右, 再根节点。  下面是代码

      public void PostOrderNonRecursion2StackV2()
            {
                Stack<TreeNode<T>> stack1 = new Stack<TreeNode<T>>();
                Stack<TreeNode<T>> stack2 = new Stack<TreeNode<T>>();
    
                TreeNode<T> current = root;
                if (current == null)
                {
                    return;
                }
    
                stack1.Push(current);
    
                while (stack1.Count != 0)
                {
                    current = stack1.Pop();
                    if (current.leftChild != null)
                    {
                        stack1.Push(current.leftChild);
                    }
    
                    if (current.rightChild != null)
                    {
                        stack1.Push(current.rightChild);
                    }
    
                    stack2.Push(current);
                }
    
                while (stack2.Count != 0)
                {
                    Console.WriteLine("value = {0}", stack2.Pop().data);
                }
            }
  • 相关阅读:
    项目中见过最好的创建树的过程
    验证Math.random()函数产生的是均匀分布的数值
    一道LINQ题,可以瞅一眼
    EntityFramework5的性能到底提高在哪里了?
    如何很好的使用Linq的Distinct方法
    EntityFramework开发三种工作流简介
    实现自动抛弃当前数据库上下文的模块。支持各种ORM框架并存
    SPS中WebPart加缓存
    交流与设计,项目管理中必备的!
    redmine 安装roadmap 插件
  • 原文地址:https://www.cnblogs.com/xuyanran/p/8287041.html
Copyright © 2011-2022 走看看