Petya has recently learned data structure named "Binary heap".

The heap he is now operating with allows the following operations:

  • put the given number into the heap;
  • get the value of the minimum element in the heap;
  • extract the minimum element from the heap;

Thus, at any moment of time the heap contains several integers (possibly none), some of them might be equal.

In order to better learn this data structure Petya took an empty heap and applied some operations above to it. Also, he carefully wrote down all the operations and their results to his event log, following the format:

  • insert x — put the element with value x in the heap;
  • getMin x — the value of the minimum element contained in the heap was equal to x;
  • removeMin — the minimum element was extracted from the heap (only one instance, if there were many).

All the operations were correct, i.e. there was at least one element in the heap each time getMin or removeMin operations were applied.

While Petya was away for a lunch, his little brother Vova came to the room, took away some of the pages from Petya's log and used them to make paper boats.

Now Vova is worried, if he made Petya's sequence of operations inconsistent. For example, if one apply operations one-by-one in the order they are written in the event log, results of getMin operations might differ from the results recorded by Petya, and some of getMin or removeMin operations may be incorrect, as the heap is empty at the moment they are applied.

Now Vova wants to add some new operation records to the event log in order to make the resulting sequence of operations correct. That is, the result of each getMin operation is equal to the result in the record, and the heap is non-empty when getMin ad removeMin are applied. Vova wants to complete this as fast as possible, as the Petya may get back at any moment. He asks you to add the least possible number of operation records to the current log. Note that arbitrary number of operations may be added at the beginning, between any two other operations, or at the end of the log.

Input

The first line of the input contains the only integer n (1 ≤ n ≤ 100 000) — the number of the records left in Petya's journal.

Each of the following n lines describe the records in the current log in the order they are applied. Format described in the statement is used. All numbers in the input are integers not exceeding 109 by their absolute value.

Output

The first line of the output should contain a single integer m — the minimum possible number of records in the modified sequence of operations.

Next m lines should contain the corrected sequence of records following the format of the input (described in the statement), one per line and in the order they are applied. All the numbers in the output should be integers not exceeding 109 by their absolute value.

Note that the input sequence of operations must be the subsequence of the output sequence.

It's guaranteed that there exists the correct answer consisting of no more than 1 000 000 operations.

Example

Input
2
insert 3
getMin 4
Output
4
insert 3
removeMin
insert 4
getMin 4
Input
4
insert 1
insert 1
removeMin
getMin 2
Output
6
insert 1
insert 1
removeMin
removeMin
insert 2
getMin 2

Note

In the first sample, after number 3 is inserted into the heap, the minimum number is 3. To make the result of the first getMin equal to 4 one should firstly remove number 3 from the heap and then add number 4 into the heap.

In the second sample case number 1 is inserted two times, so should be similarly removed twice.

题意就是模仿一个堆操作的步骤,少了几步,要让添加最少的操作 让操作正常运行。用到优先队列,不停的读,当读到一个insert操作直接将数入队就好了,没啥毛病,如果遇到了removeMin,需要判断这样的情况队列不空,那就正常,如果队列空了肯定不能正常运行,就要随便插入一个数(在这个操作之前),然后再removeMin,如果遇到的是getMin操作,有点麻烦,代码写的很详细,看代码吧,还有就是别用输入输出流,时间差的太大了 超时,换成了printf瞬间100ms。

代码:

#include <iostream>
#include <queue>
#include <cstdlib>
#include <cstdio>
#include <cstring>
#define cr "insert"
#define yc "removeMin"
#define qx "getMin"
using namespace std;
class que
{
public:
int d;
friend bool operator <(const que a,const que b)
{
return a.d>b.d;
}
}temp;
int main()
{
priority_queue<que> q;
int n,*a=new int[],k;
char b[][];
scanf("%d",&n);
for(int i=;i<n;i++)
{
scanf("%s",b[i]);
if(strcmp(b[i],yc)==)
{
if(q.empty())
{
strcpy(b[i+],yc);
strcpy(b[i],cr);
a[i]=;
i++;
n++;
}
else q.pop();
}
else
{
scanf("%d",&k);
if(strcmp(b[i],cr)==)
{
a[i]=k;
temp.d=k;
q.push(temp);
}
else
{
if(!q.empty()&&q.top().d<k)
{
while(!q.empty()&&q.top().d<k)
{
strcpy(b[i],yc);
q.pop();
i++;
n++;
}
}
if(q.empty()||q.top().d>k)
{
strcpy(b[i+],qx);;
a[i+]=k;
strcpy(b[i],cr);
a[i]=k;
i++;
n++;
temp.d=k;
q.push(temp);
}
strcpy(b[i],qx);
a[i]=k;
}
}
}
printf("%d\n",n);
for(int i=;i<n;i++)
{
if(strcmp(b[i],yc)==)printf("%s\n",b[i]);
else printf("%s %d\n",b[i],a[i]);
}
}

Heap Operations 优先队列的更多相关文章

  1. Codeforces 681C. Heap Operations 优先队列

    C. Heap Operations time limit per test:1 second memory limit per test:256 megabytes input:standard i ...

  2. Heap Operations(模拟题)

     Heap Operations time limit per test 1 second memory limit per test 256 megabytes input standard inp ...

  3. Codeforces Round #357 (Div. 2) C. Heap Operations 模拟

    C. Heap Operations 题目连接: http://www.codeforces.com/contest/681/problem/C Description Petya has recen ...

  4. CodeForces 681C Heap Operations (模拟题,优先队列)

    题意:给定 n 个按顺序的命令,但是可能有的命令不全,让你补全所有的命令,并且要求让总数最少. 析:没什么好说的,直接用优先队列模拟就行,insert,直接放入就行了,removeMin,就得判断一下 ...

  5. CodeForces 681C Heap Operations(模拟)

    比较简单的模拟,建议使用STL优先队列. 代码如下: #include<iostream> #include<cstdio> #include<cstring> # ...

  6. Codeforces Round #357 (Div. 2)C. Heap Operations

    用单调队列(从小到大),模拟一下就好了,主要是getMin比较麻烦,算了,都是模拟....也没什么好说的.. #include<cstdio> #include<map> #i ...

  7. 二叉堆(binary heap)—— 优先队列的实现

    二叉堆因为对应着一棵完全二叉树,因而可以通过线性数组的方式实现. 注意,数组第 0 个位置上的元素,作为根,还是第 1 个位置上的元素作为根? 本文给出的实现,以数组第 1 个位置上的元素作为根,则其 ...

  8. Codeforces Round #357 (Div. 2) 优先队列+模拟

    C. Heap Operations time limit per test 1 second memory limit per test 256 megabytes input standard i ...

  9. GO语言heap剖析及利用heap实现优先级队列

    GO语言heap剖析 本节内容 heap使用 heap提供的方法 heap源码剖析 利用heap实现优先级队列 1. heap使用 在go语言的标准库container中,实现了三中数据类型:heap ...

随机推荐

  1. 开关灯问题 BulbSwitch

    2018-06-17 11:54:51 开关电灯问题是一个比较经典的趣味数学题,本文中主要介绍其中的一些常见情况. 一.Bulb Switch 问题描述: 问题求解: 初始状态:off, off, o ...

  2. Python 爬虫-Scrapy框架基本使用

    2017-08-01  22:39:50 一.Scrapy爬虫的基本命令 Scrapy是为持续运行设计的专业爬虫框架,提供操作的Scrapy命令行. Scrapy命令行格式 Scrapy常用命令 采用 ...

  3. 20161210xlVBA一行数据转为四行

    Sub NextSeven_CodeFrame() '应用程序设置 Application.ScreenUpdating = False Application.DisplayAlerts = Fal ...

  4. Black Widow CodeForces - 704C (dp)

    大意: 给定一个m个bool变量的方程, 求方程解的个数 给定方程的形式类似于这样 每个括号是一个子式, 每个子式里变量数不超过2, 每个变量出现次数不超过2, 方程右侧一直是1 对每个变量出现的式子 ...

  5. Java中HashMap的实现原理

    最近面试中被问及Java中HashMap的原理,瞬间无言以对,因此痛定思痛觉得研究一番. 一.Java中的hashCode和equals 1.关于hashCode hashCode的存在主要是用于查找 ...

  6. service几种访问类型(集群外负载均衡访问LoadBalancer , 集群内访问ClusterIP,VPC内网负载均衡LoadBalancer ,集群外访问NodePort)

    一.集群外访问(负载均衡) kind: ServiceapiVersion: v1spec: ports: - protocol: TCP port: 4341 targetPort: 8080 no ...

  7. Tips for Sync Vimtex and PDF

    vimtex synctex: \lv. skim synctex: to display the TeX source line corresponding to a point in the PD ...

  8. ubuntu64位库

    安装 12.04ubuntu32位库:sudo apt-get install ia32-libs

  9. seo-摘自网友

    网页前端制作中的SEO 在SEO盛行的今天到处都在谈优化,对于网站前端制作人员来说,有几点是跟SEO相关的,也就是SEO站内优化中的一部分,下面总结几点: 1.title,<title>页 ...

  10. consumer的DubboResponseTimeoutScanTimer线程

    考虑这样一种情况,由于网络延时,consumer先抛出超时异常,一段时间后又收到了已经超时的响应,dubbo是怎么处理的? 拆分为3步看: 1. consumer的DubboResponseTimeo ...