*Leetcode_construct-binary-tree-from-inorder-and-postorder-traversal

地址： http://oj.leetcode.com/problems/construct-binary-tree-from-inorder-and-postorder-traversal/

Given inorder and postorder traversal of a tree, construct the binary tree.

Note:
You may assume that duplicates do not exist in the tree.

思路：可参考上一题：http://blog.csdn.net/flyupliu/article/details/22864057(根据先序和中序确定树)

先序和中序来确定树应该比后序和中序来确定树难一点，因为根的position要算出来。而后序和中序中，只要先创建右子树，再创建左子树，那么根就一直是后序数组的最后一个值。

这一题和上一题都跑了170ms到180ms。应该会有更简明高效的方法。

参考代码：

/**
 * Definition for binary tree
 * struct TreeNode {
 *     int val;
 *     TreeNode *left;
 *     TreeNode *right;
 *     TreeNode(int x) : val(x), left(NULL), right(NULL) {}
 * };
 */
 vector<int> invec, postvec;
 TreeNode* construct_tree(int left, int right)
 {
     if(left>right || postvec.empty())
         return NULL;
     int root_val = postvec.back();
     TreeNode* p = new TreeNode(root_val);
     postvec.pop_back();
     if(left==right)
         return p;
     int root_index = 0;
     for(; root_index < invec.size(); ++root_index)
         if(invec[root_index]==root_val)
             break;
     if(root_index==left)
     {
         p->right = construct_tree(left+1, right);
     }
     else if(root_index==right)
     {
         p->left = construct_tree(left, right-1);
     }
     else
     {
         p->right = construct_tree(root_index+1, right);
         p->left = construct_tree(left, root_index-1);
     }
     return p;
 }
class Solution {
public:
    TreeNode *buildTree(vector<int> &inorder, vector<int> &postorder) {
        if(inorder.empty())
            return NULL;
        invec = inorder;
        postvec = postorder;
        return construct_tree(0, invec.size()-1);
    }
};

  
  

   
   //SECOND TRIAL, 116ms
  
  
  
  

   
   class 
   
   Solution 
   
   {
  
  
  
  

   
   private
   
   :
  
  
  
  

   
       
   
   TreeNode 
   
   *
   
   construct
   
   (
   
   int 
   
   in
   
   [], 
   
   int 
   
   post
   
   [], 
   
   int 
   
   len
   
   )
  
  
  
  

   
       
   
   {
  
  
  
  

   
           
   
   if
   
   (
   
   len
   
   <=
   
   0
   
   )
  
  
  
  

   
               
   
   return 
   
   NULL
   
   ;
  
  
  
  

   
           
   
   int 
   
   root_val 
   
   = 
   
   post
   
   [
   
   len
   
   -
   
   1
   
   ];
  
  
  
  

   
           
   
   TreeNode
   
   * 
   
   root 
   
   = 
   
   new 
   
   TreeNode
   
   (
   
   root_val
   
   );
  
  
  
  

   
           
   
   int 
   
   offset 
   
   = 
   
   0
   
   ;
  
  
  
  

   
           
   
   while
   
   (
   
   offset
   
   <
   
   len 
   
   && 
   
   in
   
   [
   
   offset
   
   ]
   
   !=
   
   root_val
   
   )
  
  
  
  

   
               
   
   ++
   
   offset
   
   ;
  
  
  
  

   
           
   
   root
   
   ->
   
   left 
   
   = 
   
   construct
   
   (
   
   in
   
   , 
   
   post
   
   , 
   
   offset
   
   );
  
  
  
  

   
           
   
   root
   
   ->
   
   right 
   
   = 
   
   construct
   
   (
   
   in
   
   +
   
   offset
   
   +
   
   1
   
   , 
   
   post
   
   +
   
   offset
   
   , 
   
   len
   
   -
   
   1
   
   -
   
   offset
   
   );
  
  
  
  

   
           
   
   return 
   
   root
   
   ;
  
  
  
  

   
       
   
   }
  
  
  
  

   
   public
   
   :
  
  
  
  

   
       
   
   TreeNode 
   
   *
   
   buildTree
   
   (
   
   vector
   
   <
   
   int
   
   > 
   
   &
   
   inorder
   
   , 
   
   vector
   
   <
   
   int
   
   > 
   
   &
   
   postorder
   
   ) 
   
   {
  
  
  
  

   
           
   
   if
   
   (
   
   inorder
   
   .
   
   empty
   
   ())
  
  
  
  

   
               
   
   return 
   
   NULL
   
   ;
  
  
  
  

   
           
   
   int 
   
   *
   
   in 
   
   = 
   
   &
   
   inorder
   
   [
   
   0
   
   ], 
   
   *
   
   post 
   
   = 
   
   &
   
   postorder
   
   [
   
   0
   
   ];
  
  
  
  

   
           
   
   return 
   
   construct
   
   (
   
   in
   
   , 
   
   post
   
   , 
   
   inorder
   
   .
   
   size
   
   ());
  
  
  
  

   
       
   
   }
  
  
  
  

   
   };