现在的位置: 首页 > 综合 > 正文

【基础练习】链表排序,反转,划分,拷贝等

2018年04月12日 ⁄ 综合 ⁄ 共 9088字 ⁄ 字号 评论关闭
  Intersection of Two Linked Lists   28.9%
Sort List 2013-11-16 20.8% Medium
Insertion Sort List 2013-11-12 25.4% Medium
Reorder List 2013-11-02 20.5% Medium
Linked List Cycle II 2013-10-30 30.8% Medium
Linked List Cycle 2013-10-28 35.9% Medium
Copy List with Random Pointer 2013-10-03 23.6% Hard
Flatten Binary Tree to Linked List 2012-10-14 28.2% Medium
Convert Sorted List to Binary Search Tree 2012-10-02 27.4% Medium
Reverse Linked List II 2012-06-27 26.1% Medium
Partition List 2012-04-30 27.0% Medium
Remove Duplicates from Sorted List II 2012-04-22 24.8% Medium
Remove Duplicates from Sorted List 2012-04-22 34.6% Easy
Merge Two Sorted Lists 2012-03-30 33.1% Easy
Rotate List 2012-03-27 21.9% Medium
Merge k Sorted Lists 2012-02-13 21.0% Hard
Remove Nth Node From End of List 2012-01-27 28.9% Easy

Intersection of Two Linked Lists

两种情况: 存在交点,不存在(null)
先统计len1,len2. 然后判断最后一个节点的地址是否相等。不是则不存在,回null。是则先让长的走过delta步长,再同时走,直到到达相同地址,返回之。
class Solution {
public:
    ListNode *getIntersectionNode(ListNode *headA, ListNode *headB) {
        ListNode * p=headA, *pp=NULL, *q=headB, *pq=NULL, *tmp;
        int i = 0; while(p) i++, pp = p, p=p->next;
        int j = 0; while(q) j++, pq = q, q=q->next;
        if(pp!=pq) return NULL;

        if(i<j) tmp = headA, headA = headB, headB = tmp;
        for(int k = 0; k<abs(i-j); k++) headA=headA->next;
        while(headA!=headB) headA=headA->next, headB=headB->next;
        return headA;
    }
};

sort list

链表与二分法结合在一起。注意区间运算,尤其是mid。
20‘ ac  归并排序

/**
 * Definition for singly-linked list.
 * struct ListNode {
 *     int val;
 *     ListNode *next;
 *     ListNode(int x) : val(x), next(NULL) {}
 * };
 */
class Solution {
public:
    ListNode *sortList(ListNode *head) {
        if(!head) return head;
        return sort(head, NULL);
    }
    ListNode* sort(ListNode *l, ListNode *r){
        // assert(l != NULL);
        if( l ->next == r ) {
            l->next = NULL;
            return l;
        }
        
        ListNode *p=l, *p2=l;
        while(p2 != r && p2->next != r){
            p=p->next; p2=p2->next->next;
        }
        
        ListNode *a, *b, *head = NULL, *tail = NULL;
        a = sort(l, p);
        b = sort(p, r);
        while(a != NULL && b != NULL ){
            if( a->val > b->val )  {ListNode *tmp = a; a = b; b = tmp; }//@error
            if(!head) head = a;
            if(tail) tail -> next = a;
            tail = a;
            a=a->next;
        }
        if( a!= NULL ){
            if(!head) head = a;
            if(tail) tail -> next = a;
        }
        if( b!= NULL){
            if(!head) head = b;
            if(tail) tail -> next = b;
        }
        
        return head;
    }
};

上述代码要注意 

 while(p2!=r && p2->next!=r){
            p=p->next; p2=p2->next->next;
        }

这里p = mid 指向了中点(l+r)/2,注意r是不包括在内的,而 p2 指向了左闭右开 【l, r) 区间的最后一个元素,或者r。注意如果把【l,r)改写成全闭区间【L,R], 那么上面的运算的结果mid=( L  +   R+1)/2。总之mid实际上是中间偏右的一个点,不是纯正的中点。

在此基础上,就可以二分成: sort(l, r) => sort(l, p), sort(p, r)  ([l, r) => [l, (l+r)/2), [(l+r)/2, r) )

上述代码还可以优化: 对于当前链表<head, tail>,向后链入元素a,只需要以下三句:

   if(!head) head = a;
   else tail -> next = a;
   tail = a;

是不是逻辑很清晰?

我们就把上面三句话构成的append方法叫做 head-tail-append吧。

基数排序

上题也可以基数排序,毕竟保存的都是数嘛。注意弄19个链表<lists[i], tails[i]>;

从个位往高位依次排序,重复下面过程:

  • 遍历<lst, tail>中的节点it,append该节点到该位(-9 ~ 9)对应的链表<lists[i], tails[i]> 中;
  • 置<lst, tail> 为<NULL, xx>,即清空。
  • 从9到-9依次把<lists[i], tails[i]>append到<lst, tail> 中。
  • 置<lists[i], tails[i]>为<NULL, xx>,即清空。

下面是非常漂亮的一段实现代码。

ListNode *sortList(ListNode *head) {
        ListNode* lsts[19];//-9~9
        ListNode* tails[19];
        typedef ListNode* PNode;
        ListNode* lst=head,*tail;
        if(lst==NULL) return NULL;
        
        //基数排序
        int d=1;
        while(d<0xfffffff){
            for(int i=0;i<19;i++) lsts[i]=NULL;
            for(PNode it=lst;it!=NULL;it=it->next){
                int s=9+(it->val/d)%10;//0~18
                // head-tail-append 三句代码,append 节点it
                if(!lsts[s]) lsts[s]=it;
                else tails[s]->next=it;
                tails[s]=it;
            }
            lst=NULL; // <lst, tail> inited as <NULL, xx>
            for(int i=0;i<19;i++){
                if(lsts[i]){
                  // head-tail-append 三句代码再次出现
                  if(!lst) lst=lsts[i],tail=tails[i];
                  else tail->next=lsts[i],tail=tails[i];
                }
            }
            tail->next=NULL;
            
            d*=10;
        }
        return lst;
    }

把链表转成平衡二叉树

步骤和上面类似:

1 二分 dfs(l, r):  

  •  如果 l==r, 空区间 返回NULL
  •  如果 l->next == r, 返回一个节点
  •  否则 找 mid ,以mid为根节点, 二分为 dfs(l, mid), dfs(mid->next ,r)

 寻找mid的方法(下面方法找到的mid偏右)

p1 = l, p2 = l;
while(p2 !=r && p2->next !=r){
   p1 = p1->next, p2 = p2->next->next;
} //p1 将指向[l, r) 中的(l+r)/2
//一定要记得二分法中mid=(l+r)/2的含义: 
// 示意图:# # * # #, 或者 # * #
// 区间对应[l, r] 或 [l, r)

  TreeNode *convert(ListNode *l, ListNode *r){
        if(l == r) return NULL;
        if(r == l->next) return new TreeNode(l->val);
        ListNode *mid=l, *p2 = l;
        while(p2 != r && p2->next != r){
            mid = mid->next, p2 = p2->next->next;
        }
        TreeNode *lt = convert(l, mid), *rt = convert(mid->next, r);
        TreeNode *p = new TreeNode(mid->val);
        p->left = lt, p->right = rt;
        return p;
    }

或者用计数法找中点mid,则n=区间节点个数,区间[0, n-1]的中点mid的下标为( 0   +   n-1  ) / 2, 用 pm = left; for(i=0; i<mid; i++) pm = pm->next;即可得到中央节点pm

 TreeNode *dfs(ListNode *left,int n){
        if(n<=0) return NULL;
        ListNode *mid=left;
        for(int i=0;i<(n-1)/2;i++) mid=mid->next;
        TreeNode *node=new TreeNode(mid->val);
        TreeNode *pl=dfs(left,(n-1)/2),*pr=dfs(mid->next,n/2);
        node->left=pl,node->right=pr;
        return node;
    }

删除倒数第k个节点

class Solution {
public:
    ListNode *removeNthFromEnd(ListNode *head, int n) {
        ListNode *p=head,*pre = NULL, *q = head;
        int i = 1; for(; i<n && q; i++) q = q->next;
        if(i==n){
            while(q->next) pre = p, p = p->next, q = q->next;
            if(!pre){
                head = head -> next;
            }
            else pre->next = pre->next->next;
            free(p);
        }
        return head;
    }
};

或者

class Solution {
public:
    ListNode *removeNthFromEnd(ListNode *head, int n) {
        ListNode *p=head,*pre;
        int i;for(i=0;p!=NULL;i++){ p=p->next;} //i=size
        if(n>i||n<=0) return head;
        n=i-n; //point[n]
        
        pre=head;
        for(i=0;i<n-1;i++) pre=pre->next;
        if(n-1<0) {head=head->next; return head;}
        pre->next=pre->next->next;
        return head;
    }
};

划分List (小于x | 大于等于x)

class Solution {
public:
    ListNode *partition(ListNode *head, int x) {
        ListNode *p=head,*pre=NULL;
        while(p!=NULL&&p->val<x){ pre=p,p=p->next;}
        if(p!=NULL){//p->val>=x
            ListNode *q=p->next,*preq=p;
            while(q!=NULL){
                if(q->val<x){
                    //cut q and insert before p
                    preq->next=q->next;
                    
                    if(pre!=NULL) {pre->next=q;}
                    else {head=q;}//head change
                    q->next=p;
                    pre=q;
                }
                preq=q,q=q->next;//@attention: not good 
            }
        }
        return head;
    }
};

或者

class Solution {
public:
    ListNode *partition(ListNode *head, int x) {
        ListNode * p = head, *pp = NULL;
        while(p && p->val < x) pp = p, p = p->next;
        if(p){
           ListNode * q = p->next, * pq = p;
           while(q){
            if( q->val < x ) 
            {
               pq -> next = pq->next->next;
               if(!pp) head = q;
               else pp -> next = q;
               pp = q;
               pp -> next = p;
               q = pq -> next;//@important
            } 
            else pq = q, q = q->next;
          }
        }
        return head;
    }
};

copy list with random pointer

整体上分为三步骤:

第一步对每个节点拷贝一个新的跟在后面,

第二步非常重要:在这里而不是在后面修改新节点的random指针,并且注意“空指针”

第三步,拆分原有链表和新增链表(用p1,q1分别表示新增表头尾指针即可)。用到了上述的head-tail-append三句代码。

/**
 * Definition for singly-linked list with a random pointer.
 * struct RandomListNode {
 *     int label;
 *     RandomListNode *next, *random;
 *     RandomListNode(int x) : label(x), next(NULL), random(NULL) {}
 * };
 */
class Solution {
public:
    RandomListNode *copyRandomList(RandomListNode *head) {
        RandomListNode * p = head, *np;
        while(p){
            RandomListNode * t = new RandomListNode(p->label);
            t->random = p->random;
            t->next = p->next;
            p->next = t;
            p = t->next;
        }
        //@error: should change random below, not later
        p = head;
        while(p){
            np = p->next;
            if(np->random) np->random = np->random->next;
            p = np->next;
        }
        
        p = head;
        RandomListNode *p1=NULL, *q1;
        while(p){
            np = p->next;
           // if(np->random) np -> random = np->random->next; //@error: should be changed earlier
            if(p1) q1->next = np;
            else p1 = np;
            q1 = np;
            p -> next = np->next;
            p = p -> next;
        }
        return p1;
    }
};

链表插入排序

链表插入排序用到了    链表的遍历 和  插入(区分头插入和中间插入),  

步骤:

1  <head, .... NULL> 为已经排序的部分, 初始化为<head, NULL>
    <p, ...NULL> 为待插入的部分,初始化为<head->next, NULL>
2 每次在<head,.. .NULL> 中寻找p的插入点<pre, q=pre->next>, 
  • 将p->next 置为NULL,准备插入(此外np记录下一个p)
  •  如果pre==NULL,说明q=head, p将插入到<head, NULL>之前,于是头部插入, p替代head (p->next=head, head=p;)
  • 否则, 在<head, NULL> 中间或末尾插入,如果在末尾则q为NULL (pre->next = p, p->next = q; )
class Solution {
public:
    ListNode *insertionSortList(ListNode *head) {
        if(!head) return head;
        ListNode *q =NULL, *pre = NULL, *p = head->next, *np;
        head -> next = NULL;
        while(p){
            np = p->next;
            p->next = NULL; // cut p
            
            pre = NULL, q = head;
            while(q && q->val <= p->val) pre = q, q = q->next;
            if(!pre) p->next = head, head = p;
            else  p -> next = pre -> next, pre->next = p;
            
            p = np;
        }
        return head;
    }
};

Reorder List

只想到了O(n)空间复杂度的解法。有O(1)空间的解法吗?

void reorderList(ListNode *head) {
        if(!head) return;
        vector<ListNode *> vec;
        ListNode * p =head;
        while(p) vec.push_back(p), p = p->next;
        int n = vec.size();
        for(int i = 0; i<=(n-1)/2; i++){
            if(i>0) vec[n-i]->next = vec[i];
            vec[i]->next = vec[n-1-i];
            vec[n-1-i]->next = NULL;
        }
    }

Remove Duplicate 2

class Solution {
public:
    ListNode *deleteDuplicates(ListNode *head) {
        ListNode * pre = NULL, * p = head, *np;
        head = NULL; // <head, pre> inited to <NULL, xx>
        while(p){
            ListNode * q = p->next, *tmp;
            while(q && q->val == p->val) q = q->next;
            // append p to <head, pre>
            if(q==p->next){ 
                p->next = NULL;
                if(!pre) head = p;
                else pre->next = p;
                pre = p;
            }
            //skip duplicated p
            else{
                while(p!=q) {
                    ListNode *tmp = p->next;
                    free(p);
                    p = tmp;
                }
            }
            //next p
            p = q;
        }
        return head;
    }
};

或者 

保持左边<head, pre>和右边的p的链接。

ListNode *deleteDuplicates(ListNode *head) {
        ListNode *p=head,*pre=NULL;
        while(p!=NULL&&p->next!=NULL){
           if(p->val==p->next->val){
              int val=p->val; ListNode *next;
              while(p!=NULL&&p->val==val){
                next=p->next;
                delete p;
                p=next;
              }
              // <head, pre> link to next p
              if(pre!=NULL) pre->next=p;
              else { head=p; }
           }
           else{
              // p append to <head, pre>
               pre=p,p=p->next;
           }
        }
        return head;
    }
};

swap pairs

注意np,pp分别代表什么

 ListNode *swapPairs(ListNode *head) {
        ListNode * p = head, *q = p, *np, *pp = NULL;
        while(p && p->next){
            np = p->next->next;
            q = p->next;
            q -> next = p;
            // link last section with q
            if(pp) pp -> next = q;
            // first section, then head change
            else head = q;
            pp = p, p = np;
        }
        // link last section with p=NULL or single element
        if(pp) pp->next = p;
        return head;
    }

rotate list

这个题目比较坑爹的地方在于k的大小可以超过n, 

记得shift k相当于把下标为(n-k%n)%n的元素给作为首节点,就OK了

 ListNode *rotateRight(ListNode *head, int k) {
        int n =0;
        ListNode * p = head, * pre = NULL, *tail = NULL;
        while(p){ n++; tail = p, p = p->next; }
        if(n==0) return head;
        k = (n - k%n)%n;
        if(k==0) return head;
        p = head;
        for(int i=0; i<k; i++){ pre = p, p = p->next; }
        tail -> next = head, head = p, pre -> next = NULL;
        return head;
    }

add two numbers

非常优美的,从第一个节点汪最后一个节点计算和append:

    ListNode *addTwoNumbers(ListNode *l1, ListNode *l2) {
        // @error: no need to reverse l1 or l2
        /*ListNode * p = l2, *pp =NULL;
        while(p){
            ListNode *np = p->next;
            p->next = pp;
            pp = p, p = np;
        }
        l2 = pp;*/
        
        ListNode * l3 = NULL, *tail;//head
        int c =0;
        while(l1 || l2 || c){
            int a= l1? l1->val:0;
            int b = l2? l2->val:0;
            if(l1) l1 = l1->next;
            if(l2) l2 = l2->next;
            
            int t = a + b + c;
            if(t>=10) t%=10, c=1;
            else c = 0;
            ListNode * tmp = new ListNode(t);
            //tmp->next = l3, l3 = tmp; //@error: should not insert in the head of <l3, xx>, because head is the lowest bit
            if(!l3) l3 = tmp; //@attention: should append tmp to <l3, tail>
            else tail->next = tmp;
            tail = tmp;
        }
        return l3;
    }

抱歉!评论已关闭.