leetcode-CloneGraph

题目地址：https://leetcode.com/problems/clone-graph/

---

question:

---

my thinking:
其实这个题目是基于图的遍历的，不过要在遍历的过程中还要对节点进行深拷贝，我们知道图的遍历一般是有两种方法的：BFS（breadth-first-search）和DFS(depth-first-search)，这里我是采用的BFS算法，结合queue存储要访问的节点（和最基本的BFS中使用的queue是一样的作用）和map存储原节点指针和对应新图指针的映射（这样的话就能知道哪个节点有没有被复制过并能以O（1）的时间复杂度找到复制过的节点指针）

---

my answer:

C++
/**
 * Definition for undirected graph.
 * struct UndirectedGraphNode {
 *     int label;
 *     vector<UndirectedGraphNode *> neighbors;
 *     UndirectedGraphNode(int x) : label(x) {};
 * };
 */
class Solution {
public: 
    UndirectedGraphNode *cloneGraph(UndirectedGraphNode *node) 
    {
        queue<UndirectedGraphNode *> nodesQueue;
        map<UndirectedGraphNode *,UndirectedGraphNode *> nodesMap;
        if(node==NULL)
        {
            return NULL;
        }
        nodesQueue.push(node);
        UndirectedGraphNode *newRoot = new UndirectedGraphNode(node->label);
        nodesMap.insert(pair<UndirectedGraphNode *,UndirectedGraphNode *>(node,newRoot));

        while(!nodesQueue.empty())
        {
            UndirectedGraphNode *currNode = nodesQueue.front();
            nodesQueue.pop();
            for(vector<UndirectedGraphNode *>::iterator iter=currNode->neighbors.begin();iter!=currNode->neighbors.end();++iter)
            {
                if(nodesMap.count(*iter)==0)//还没有被遍历过
                {
                    UndirectedGraphNode *newNode = new UndirectedGraphNode((*iter)->label);
                    nodesMap.insert(pair<UndirectedGraphNode *,UndirectedGraphNode *>(*iter,newNode));
                    ((nodesMap.find(currNode))->second)->neighbors.push_back(newNode);
                    nodesQueue.push((*iter));
                }
                else
                {
                     ((nodesMap.find(currNode))->second)->neighbors.push_back((nodesMap.find(*iter)->second));
                }
            }
        }
        return newRoot;
    }
};

---

code explation:
就像前面说的一样，主要是基本的BFS遍历加上深拷贝节点值建立节点之间的连通关系，nodesMap存储的是原节点指针和复制后节点指针之间的映射，在一开始把头节点和复制后的头节点组成pair插入到nodesMap中:UndirectedGraphNode *newRoot = new UndirectedGraphNode(node->label);
nodesMap.insert(pair<UndirectedGraphNode *,UndirectedGraphNode *>(node,newRoot));
然后在BFS的遍历过程中查看某个节点是不是被复制过（=是否被遍历过），如果还未被复制过的话，复制该节点，插入到nodesMap中建立映射，并在当前访问节点所对应的新拷贝节点的neighbors邻节点列表中加入新拷贝的节点:
UndirectedGraphNode *newNode = new UndirectedGraphNode((*iter)->label);
nodesMap.insert(pair<UndirectedGraphNode *,UndirectedGraphNode *>(*iter,newNode));
((nodesMap.find(currNode))->second)->neighbors.push_back(newNode);
如果已经被复制过的话，直接在当前访问的节点所对应的新拷贝节点的neighbors中加入之前存储的映射节点:
((nodesMap.find(currNode))->second)->neighbors.push_back((nodesMap.find(*iter)->second));