• Sort List[leetcode] 由归并排序的递归和循环,到本题的两种解法


    归并排序能够有两种思路----top-down 和 bottom-up

    top-down:

    递归实现,将数组分成两半。分别处理。再合并。

    伪代码例如以下

    split ( A[], l, r)
    {
    	if ( r - l < 2) return;
    	m = (r + l) / 2;
    	split ( A, l, m); //split A[l…m-1]
    	split ( A, m, r); //split A[m…r-1]
    	merge ( A, l, m, e); //merge A[l…m-1] and A[m…e-1]
    }

    bottom-up:

    循环实现。将数组看做n个长度为1的组数组。

    用width控制每次merge的子数组长度。width每次翻倍

    伪代码例如以下

    sort ( A[], n)
    {
    	for (width = 1; width < n; width *= 2)
    	{
    		for (i = 0; i < n; i += 2 * width)
    		{
    			merge(A, i, min(i + width, n), min(i + 2 * width, n));
    		}
    	}
    }

    Sort list中使用链表。不能在O(1)的时间内訪问随意节点,同一时候注意要处理尾部节点的next,置为NULL

    和上面的伪代码类似,首先实现merge函数:

        ListNode * merge(ListNode * h1, int s1, ListNode * h2, int s2)
        {
            if (h2 == NULL) return h1;
            ListNode * h;
            if (h1->val < h2->val)
                h = advance(h1, s1);
            else
                h = advance(h2, s2);
            ListNode * cur = h;
            while (s1 && s2)
            {
                if (h1->val < h2->val)
                    cur->next = advance(h1, s1);
                else
                    cur->next = advance(h2, s2);
                cur = cur->next;
            }
            if (s1)
            {
                cur->next = h1;
                while(s1) advance(h1, s1);
            }
            if (s2)
            {
                cur->next = h2;
                while(s2) advance(h2, s2);
            }
            return h;
        }
        
        ListNode * advance(ListNode * (& n), int & size)
        {
            ListNode * temp = n;
            if (size == 1)  n->next = NULL;
            n = n->next;
            size--;
            return temp;
        }
    


    同一时候实现工具函数,訪问任何位置节点

    ListNode * getNode(ListNode * head, int len)
        {
            while (len -- && head) head = head->next;
            return head;
        }

    循环版本号主函数例如以下:

    ListNode *sortList(ListNode *head) {
            ListNode * cur = head;
            int size = 0;
            while (cur)
            {
                size ++;
                cur = cur->next;
            }
            
            ListNode * pre;
            for (int w = 1; w <= size; w *= 2)
            {
                cur = head;
                for (int i = 0; i < size; i+= w*2)
                {
                    ListNode * h1 = cur, * h2 = getNode(cur, w), * next = getNode(cur, 2 * w);
                    cur = merge(h1, min(w, size - i), h2, min(w, size - i - w));
                    if (i == 0)     
                        head = cur;
                    else            
                        pre->next = cur;
                    pre = getNode(cur, min(2 * w, size - i) - 1);
                    cur = next;
                }
            }
            return head;
        }

    递归版本号主函数例如以下:

        ListNode *sortList(ListNode *head) {
            ListNode * cur = head;
            int size = 0;
            while (cur)
            {
                size ++;
                cur = cur->next;
            }
            return sort(head, size - size / 2, getNode(head, size - size / 2), size / 2);
        }
        
        ListNode * sort(ListNode * h1, int s1, ListNode * h2, int s2)
        {
            if (s1 == 0) return h2;
            if (s2 == 0) return h1;
            h1 = sort(h1, s1 - s1 / 2, getNode(h1, s1 - s1 / 2), s1 / 2);
            h2 = sort(h2, s2 - s2 / 2, getNode(h2, s2 - s2 / 2), s2 / 2);
            return merge(h1, s1, h2, s2);
        }



  • 相关阅读:
    菜鸟系列Golang学习 — 协程
    菜鸟系列Golang学习 — 切片
    菜鸟系列Golang学习 — 数组
    菜鸟系列Fabric —— Fabric-CA
    菜鸟系列计算机网络——TCP和UDP协议
    菜鸟系列Fabric源码学习 — kafka共识机制
    菜鸟系列Fabric源码学习 — 背书节点和链码容器交互
    Amazon's Dynamo 论文中文版
    3、递归(动态规划思想)
    2、链表(python实现)
  • 原文地址:https://www.cnblogs.com/blfbuaa/p/6743700.html
Copyright © 2020-2023  润新知