基本上官方的思路了,一开始没用这个方法是觉得直接接到新的链表里,后面会拖着一串,但这个全部的节点都会包括到,其实也没什么问题。另外感觉还是尽量减少特殊情况的讨论,比较费调试时间。只在NULL节点无法再判断val了特殊算一下就OK了。

/**
 * struct ListNode {
 *	int val;
 *	struct ListNode *next;
 * };
 */
/**
 * 代码中的类名、方法名、参数名已经指定,请勿修改,直接返回方法规定的值即可
 *
 * 
 * @param pHead1 ListNode类 
 * @param pHead2 ListNode类 
 * @return ListNode类
 */
#include <stdlib.h>
struct ListNode* Merge(struct ListNode* pHead1, struct ListNode* pHead2 ) {
    // write code here
    struct ListNode* new_head = (struct ListNode*)malloc(sizeof(struct ListNode));
    new_head->val = 0;
    struct ListNode* cur = new_head;
    while(pHead1 != NULL && pHead2 != NULL)
    {
        if(pHead1->val <= pHead2->val)
        {
            cur->next = pHead1;
            cur = cur->next;
            pHead1 = pHead1->next; 
        }
        else if(pHead1->val > pHead2->val)
        {
            cur->next = pHead2;
            cur = cur->next;
            pHead2 = pHead2->next; 
        }
    }
    if(pHead1 == NULL)
    {
        cur->next = pHead2;
        return new_head->next;
    }
    if(pHead2 == NULL)
    {
        cur->next = pHead1;
        return new_head->next;
    }
    return new_head->next;

}