统计一篇文章中各英语单词出现的频数

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/HSH205572/article/details/82872359

package com.icinfo;

import java.io.*;
import java.util.*;

/**
 * 统计一个文件中各词出现的频率，并打印
 */
public class FileWordCount {
    // 使用HashMap来存储单词的频率
    Map<String, Integer> wordCount = new HashMap<>();

    public static void main(String[] args) {
        HashMap<String, Integer> map = (HashMap<String, Integer>) new FileWordCount()
                .wordCount("C:/Users/hzhb/Desktop/test.txt");

        // 自定义排序
        List<Map.Entry<String, Integer>> list = new LinkedList<>();
        list.addAll(map.entrySet());
        list.sort(Comparator.comparingInt(e -> e.getValue()));
        list.forEach(System.out::println);
    }

    /**
     * @param fileName 文件名（将英文文章复制到一个文件中去）
     */
    public Map<String, Integer> wordCount(String fileName) {
        File file = new File(fileName);
        FileInputStream fis = null;
        try {
            fis = new FileInputStream(file);
        } catch (FileNotFoundException e) {
            System.out.println("文件不存在!");
        }

        BufferedReader bufr = new BufferedReader(new InputStreamReader(fis));
        String s;
        try {
            while ((s = bufr.readLine()) != null) {
                // 移除字符串的前导空白和后尾部空白
                s = s.trim();
                // 正则表达式：以非字母或者是数字为分隔符，进行分割
                // 英文单词以空格为分隔符，将单词分隔
                String[] str = s.split("(\\s+\\W+)|[\\s+\\W+]");
                for (int i = 0; i < str.length; i++) {
                    //并将所有大写字母转换为小写
                    String currentStr = str[i].toLowerCase();
                    // 如果HashMap中已有该值,将值加1
                    if (wordCount.containsKey(currentStr)) {
                        wordCount.put(currentStr, wordCount.get(currentStr) + 1);
                    } else {
                        // 默认初始化该单词的出现次数为1
                        wordCount.put(currentStr, 1);
                    }
                }
            }
        } catch (IOException e) {
            e.printStackTrace();
        }

        // 移除HashMap中的""空字符串
        wordCount.remove("");
        return wordCount;
    }
}