多线程单词统计

#include<iostream>
#include<algorithm>
#include<cstdio>
#include<stdlib.h>
#include<pthread.h>
#include<map>
#include<string>
#include<vector>
using namespace std;
void *count(void *filename); //统计单词数
void allcount_printf(); //输出结果
map<string,int> word_count;	//存放单词对应出现次数
map<string,int>::iterator iter; //迭代器
bool cmp(const pair<string,int> &x,const pair<string,int> &y)
{
	return x.second>y.second;
}
pthread_mutex_t mutex;
int main(int argc,char *argv[]){
	clock_t start,finish;
	double totaltime;
	start=clock();
	pthread_t tid[100];
	//初始化互斥锁
	pthread_mutex_init(&mutex,NULL);
	//创建子线程
	for(int i=1;i<argc;i++)
		pthread_create(&tid[i],NULL,count,(void *)argv[i]);
	//使主线程等待子线程
	for(int i=1;i<argc;i++)
		pthread_join(tid[i],NULL);
	//销毁互斥锁
	pthread_mutex_destroy(&mutex);
	//输出一共有多少不同的单词
	int cnt=word_count.size();
	printf("total number of words:%d\n",cnt);
	//按出现频率大小输出所有单词
	allcount_printf();
	finish=clock();
   	totaltime=(double)(finish-start)/CLOCKS_PER_SEC;
   	cout<<"\n多线程程序的运行时间为"<<totaltime<<"秒！"<<endl;
}
void *count(void *filename)
{
	char *fname=(char *)filename;
	FILE *fp=NULL;
	char tmp[100]; 
	//打开并读取文件
	if((fp=fopen(fname,"r"))==NULL){
		cout<<"error"<<endl;
	}
	else 
	{
		while(fscanf(fp,"%s",tmp)!=EOF){
			string str=tmp;
			//将所有单词都变成小写 避免因大小写区分相同单词
			transform(str.begin(),str.end(),str.begin(),::tolower);
			//去掉字符串中的标点符号
			string::iterator pos=str.begin();
			while(pos!=str.end())
			{
				if(ispunct(*pos))
					str.erase(pos);
				else
					pos++;
			}
			//判断map中是否为空或者读入的单词是否已经插入
			if(word_count.empty()){
				pthread_mutex_lock(&mutex);
				word_count.insert(pair<string,int>(str,1));
				pthread_mutex_unlock(&mutex);
			}
			else if(word_count.find(str)==word_count.end()){
				pthread_mutex_lock(&mutex);
				word_count.insert(pair<string,int>(str,1));
				pthread_mutex_unlock(&mutex);
			}
			else{
				pthread_mutex_lock(&mutex);
				word_count[str]++;
				pthread_mutex_unlock(&mutex);
			}
		}
	}
}
void allcount_printf()
{
	//创建一个vector实现map按value进行降序排序
	vector< pair<string,int> >word;
	for(iter=word_count.begin();iter!=word_count.end();iter++)
		word.push_back(pair<string,int>(iter->first,iter->second));
	sort(word.begin(),word.end(),cmp);
	//输出TOP10热词
	printf("TOP10热词为：\n");
	for(int i=0;i<10;i++)
		cout<<word[i].first<<":"<<word[i].second<<endl;
	//按出现频率大小降序输出每个单词
	printf("各单词出现频率为：\n");
	for(int i=0;i!=word.size();i++)
		cout<<word[i].first<<":"<<word[i].second<<endl;
}

运行结果：