统计单词出现频率及排序 从单机到多机合作

本文是学习 多线程服务端编程的练习

书籍作者陈硕的博客也有提到这个题目

http://blog.csdn.net/solstice/article/details/8497475

 

第一个层次很简单 单机 一个小文件 读进来进行处理 然后对每个单词进行统计排序 记录每个单词出现频率

#include <algorithm>
#include <iostream>
#include <unordered_map>
#include <vector>
#include <sstream>
#include <fstream>
 
using namespace std;
 
std::unordered_map<std::string, int> mapStringFrequent;
 
 
void SplitStrig(std::string& line, std::vector<std::string>& vecString) {
    typedef std::string::size_type strSize;
    strSize i = 0;
    while (i != line.size()) {
        while (i != line.size() && isspace(line[i]))
            ++i;
        strSize j = i;
        while (j != line.size() && !isspace(line[j]))
            ++j;
        if (i != j) {
            vecString.push_back(line.substr(i, j - i));
            i = j;
        }
    }
}
 
bool CheckWord(std::string s) {
    typedef std::string::size_type strSize;
    strSize i = 0;
    while (i != s.size()) {
        if (isalpha(s[i])) {
            i++;
        }
        else {
            return false;
        }
    }
    return true;
}
 
 
 
void VecInputMap(std::vector<std::string>& vecString, std::unordered_map<std::string, int>& mapStringFrequent) {
    for (auto it : vecString) {
        if (CheckWord(it)) {
            transform(it.begin(), it.end(), it.begin(), ::tolower);
            mapStringFrequent[it]++;
        }
    }
}
 
 
bool ReadFile(const string& fileName) {
    bool ret = false;
    std::ifstream in(fileName);
    if (in.bad() || !in) {
        return ret;
    }
    std::string line;
    std::vector<std::string> vecString;
    try {
        while (getline(in, line))
        {
            //std::cout << line <<  std::endl;
            SplitStrig(line, vecString);
        }
 
        VecInputMap(vecString, mapStringFrequent);
    }
    catch (std::exception& e) {
        std::cout << line << std::endl;
        in.close();
        std::cerr << e.what() << std::endl;
        return ret;
    }
 
    in.close();
    ret = true;
    return ret;
}
 
 
void WriteToFile(const string& fileName, std::vector<std::pair<int, std::string>>& vecWordFreq) {
    std::ofstream out(fileName);
    if (out.bad() || !out) {
        return;
    }
 
    for (auto& it : vecWordFreq) {
        out << it.second << '\t' << it.first << std::endl;
    }
    out.close();
 
}
//============================================================
 
int main()
{
    bool ret = ReadFile("1.txt");
    if (!ret)
        return -1;
 
    //ret = ReadFile("2.txt");
    //if (!ret)
    //  return -1;
    //ret = ReadFile("3.txt");
    //if (!ret)
    //  return -1;
    //ret = ReadFile("4.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("5.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("6.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("7.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("8.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("9.txt");
    //if (!ret)
    //  return -1;
 
    //ret = ReadFile("10.txt");
    //if (!ret)
    //  return -1;
 
    std::vector<std::pair<int, std::string>> freq;
    freq.reserve(mapStringFrequent.size());
 
    for (auto& it : mapStringFrequent){
        freq.push_back(make_pair(it.second, it.first));
    }
 
    std::sort(freq.begin(), freq.end(), [](const std::pair<int, std::string>& lhs,  // const auto& lhs in C++14
            const std::pair<int, std::string>& rhs) {
        return lhs.first > rhs.first;
    });
 
    //for (auto it : freq) {
    //  std::cout << it.first << '\t' << it.second << '\n';
    //}
 
    WriteToFile("freqResult.txt", freq);
    return 0;
}

第二个层次 就是文件较大 单词量较多 如果一次性读入并使用vector容器存放 以及使用hash容器进行统计频率 会出现内存等资源不足现象

那么就依次读取进内存 将解析的单词存放进vector中，当vector中的容量到达一个阈值 将vector中的单词放进map容器中统计单词出现频率 vector容器清空

但是map容器有多个 各个单词存放进那个map容器根据hash函数决定  

当map容器的容量达到阈值后 写入文件 map容器清空

这样我们就得到多个记录单词出现频率的文本 供后继步骤分析合并 但是一个单词的出现频率肯定只出现在一个记录文本中(因为相同的单词具有相同的hash值 对应同一个记录文本)

如图

 

代码如下(记录合并代码后继完成)

#include <algorithm>
#include <iostream>
#include <unordered_map>
#include <vector>
#include <sstream>
#include <fstream>
 
using namespace std;
 
#define BUCKET_NUM  10
std::unordered_map<std::string, int> mapStringFrequent[BUCKET_NUM];
std::ofstream out[BUCKET_NUM];
 
 
#define MAX_WORD_SIZE   1024*1024
 
 
void SplitStrig(std::string& line, std::vector<std::string>& vecString) {
    typedef std::string::size_type strSize;
    strSize i = 0;
    while (i != line.size()) {
        while (i != line.size() && isspace(line[i]))
            ++i;
        strSize j = i;
        while (j != line.size() && !isspace(line[j]))
            ++j;
        if (i != j) {
            vecString.push_back(line.substr(i, j - i));
            i = j;
        }
    }
}
 
bool CheckWord(std::string s) {
    typedef std::string::size_type strSize;
    strSize i = 0;
    while (i != s.size()) {
        if (isalpha(s[i])) {
            i++;
        }
        else {
            return false;
        }
    }
    return true;
}
 
void VecInputMap(std::vector<std::string>& vecString, std::unordered_map<std::string, int> mapStringFrequent[]) {
    for (auto it : vecString) {
        if (CheckWord(it)) {
            transform(it.begin(), it.end(), it.begin(), ::tolower);
            int idx = std::hash<string>()(it) % BUCKET_NUM;
            mapStringFrequent[idx][it]++;
            if (mapStringFrequent[idx].size() >= MAX_WORD_SIZE) {
                for (auto& it : mapStringFrequent[idx]) {
                    out[idx] << it.second << '\t' << it.first << std::endl;
                }
                mapStringFrequent[idx].clear();
            }
        }
    }
}
 
bool ReadFile(const string& fileName) {
    bool ret = false;
    std::ifstream in(fileName);
    if (in.bad() || !in) {
        return ret;
    }
    std::string line;
    std::vector<std::string> vecString;
 
    try {
        while (getline(in, line))
        {
            SplitStrig(line, vecString);
            //std::cout << line << std::endl;
            if (vecString.size() >= MAX_WORD_SIZE) {
                //std::cout << "too many words" << std::endl;
                VecInputMap(vecString, mapStringFrequent);
                vecString.clear();
            }
        }
 
        for (int i = 0; i < BUCKET_NUM; i++) {
            for (auto& it : mapStringFrequent[i]) {
                out[i] << it.second << '\t' << it.first << std::endl;
            }
            mapStringFrequent[i].clear();
        }
 
    }
    catch (std::exception& e) {
        in.close();
        std::cerr << e.what() << std::endl;
        return ret;
    }
 
    std::cout << vecString.size() << std::endl;
    ret = true;
    return ret;
}
 
int main()
{
    std::string fileName = "0bucket.txt";
    for (int i = 0; i < BUCKET_NUM; i++) {
        out[i].close();
        out[i].open(fileName);
        fileName[0]++;
    }
 
    ReadFile("LargeFile.txt");
 
     
    return 0;
}

运行效果如图

运行前

运行后

合并步骤如图:

合并代码如下

// WordFrequentMerge.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"
#include <algorithm>
#include <iostream>
#include <unordered_map>
#include <deque>
#include <vector>
#include <sstream>
#include <fstream>

using namespace std;


#define DEFAULT_FILE_NAME	"0bucket.txt"
#define BUCKET_NUM	10
std::ifstream in[BUCKET_NUM];
std::deque<std::pair<int, std::string>> deqFreqRecord[BUCKET_NUM];



void OpenFile() {
	std::string fileName = "0bucket.txt";
	for (int i = 0; i < BUCKET_NUM; i++) {
		in[i].close();
		in[i].open(fileName);
		if (in[i].bad() || !in[i]) {
			std::cerr << "open file error.exit!!" << std::endl;
			return;
		}
		fileName[0]++;
	}
}

void ReadSortEle_(std::ifstream& in, std::deque<std::pair<int, std::string>>& d) {
	std::string line,word;
	int freq;
	try {
		while (getline(in, line)) {
			std::string::size_type pos = line.find('\t');
			if (pos != std::string::npos) {
				freq = std::stoi(line.substr(0, pos));
				word = line.substr(pos + 1);
				//std::cout << word << "\t" << freq << std::endl;
				d.push_back(std::make_pair(freq, word));
			}
		}
	}
	catch (std::exception& e) {
		std::cout << e.what() << std::endl;
		return;
	}

	std::sort(d.begin(),d.end(), [](const std::pair<int, std::string>& lhs,  
		const std::pair<int, std::string>& rhs) {
		return lhs.first > rhs.first;
	});
}


void ReadSortEle(std::ifstream in[], std::deque<std::pair<int, std::string>> d[]) {
	for (int i = 0; i < BUCKET_NUM; i++) {
		ReadSortEle_(in[i],d[i]);
	}
}

bool HeapCompareFunc(std::pair<int, std::string>& l,
	std::pair<int, std::string>& r) {
	return l.first < r.first;
}

void MergeFrequent(std::deque<std::pair<int, std::string>> d[]) {
	vector<std::pair<int, std::string>> vecHeap;
	for (int i = 0; i < BUCKET_NUM; i++) {
		if (!d[i].empty()) {
			vecHeap.push_back(d[i].front());
			d[i].pop_front();
		}
	}

	std::make_heap(vecHeap.begin(), vecHeap.end(), HeapCompareFunc);

	while (!vecHeap.empty()) {
		std::pop_heap(vecHeap.begin(), vecHeap.end(), HeapCompareFunc);
		std::pair<int, std::string> p = vecHeap.back();
		vecHeap.pop_back();

		std::cout << p.second << "\t" << p.first << std::endl;

		int idx = std::hash<string>()(p.second) % BUCKET_NUM;
		
		if (!d[idx].empty()) {
			vecHeap.push_back(d[idx].front());
			std::push_heap(vecHeap.begin(), vecHeap.end(), HeapCompareFunc);
			d[idx].pop_front();
		}
	}

}


int main()
{
	OpenFile();
	ReadSortEle(in, deqFreqRecord);
	MergeFrequent(deqFreqRecord);

    return 0;
}

代码还有诸多待优化地方

比如读入单词后 另开线程处理而不是等待处理完毕后再读取单词

比如数据的传递 考虑传递指针和起始位置 而不是传递拷贝

技术博客 http://blog.csdn.net/stecdeng 技术交流群 群号码：324164944 欢迎c c++ windows驱动爱好者 服务器程序员沟通交流