软件工程作业二：WordCount实现与改进

Posted 2020-10-07 ReverseAC

首先，我在网上找了一个WordCount的源码，在源码上进行修改

原文博主：波涛汹涌

原文链接：http://blog.sina.com.cn/s/blog_51335a000101ouzf.html

原文代码：

wordCountDemo

原文的代码实现读取文件，统计所有出现过的单词（不重复），并统计打印每个单词出现的次数。代码导入包时有两个错误。且打印存在重复，输出不好看。

修改后的代码：

  1 import java.io.BufferedReader;
  2 import java.io.BufferedWriter;
  3 import java.io.File;
  4 import java.io.FileInputStream;
  5 import java.io.FileReader;
  6 import java.io.FileWriter;
  7 import java.io.IOException;
  8 import java.io.InputStreamReader;
  9 import java.util.ArrayList;
 10 import java.util.Collections;
 11 import java.util.HashMap;
 12 import java.util.HashSet;
 13 import java.util.List;
 14 import java.util.Map;
 15 import java.util.Set;
 16 import java.util.TreeMap;
 17 
 18 /**
 19  * 
 20  * @author 张军洪 
 21  * original author：波涛汹涌
 22  * http://blog.sina.com.cn/s/blog_51335a000101ouzf.html
 23  *
 24  */
 25 
 26 public class WordCount {
 27     private static String path_src = "C:/Users/张军洪/Desktop/shenma.txt";
 28     private static String path_result = "C:/Users/张军洪/Desktop/shenma_result.txt";
 29     private static BufferedReader br = null;
 30     private static BufferedWriter bw = null;
 31     private static String line_current = null;
 32     private static String[] words = null;
 33     private static List<String> word_list = new ArrayList<String>();  //所有出现的单词(不重复)
 34     static int countWord=0;
 35     static int countChar=0;
 36     static int countLine=0;
 37     static String s="";
 38     static String strCount="";
 39     
 40     public static void main(String[] args) {
 41 
 42         File file = new File(path_src); // 读取文件
 43 
 44         // 错误处理
 45         if (!file.exists()) {
 46             System.out.println("file " + file + " is not existed, exit");
 47             return;
 48         }
 49         try {
 50             br = new BufferedReader(new FileReader(file.getPath()));
 51             line_current = br.readLine();
 52             while (line_current != null) {
 53                 words = line_current.split(" |,|\\.");
 54                 for (String s : words) {
 55                     if (!s.equals(""))
 56                         word_list.add(s);
 57                 }
 58 
 59                 line_current = br.readLine();
 60             }
 61 
 62             // 打印单词
 63             for (String temp : word_list) {
 64                 System.out.println(temp);
 65             }
 66              br = new BufferedReader(new InputStreamReader(new FileInputStream(new File(path_src))));
 67              while((s=br.readLine())!=null)
 68               {
 69                  s=s+" ";
 70                  strCount+=s;
 71                countLine++;
 72               }
 73              for(int i=0;i<strCount.split(" ").length;i++){
 74                  if(!strCount.split(" ")[i].equals(" "))
 75                      countWord++;
 76                  countChar+= strCount.split(" ")[i].length();
 77              } 
 78              System.out.println();
 79 //             System.out.println("单词数："+countWord);
 80 //             System.out.println("字符数："+countChar);
 81 //             System.out.println("行数："+countLine);
 82 //             System.out.println();
 83 
 84             // HashSet
 85             Set<String> hashSet = new HashSet<String>(word_list);
 86 //            for (String str : hashSet) {
 87 //                System.out.println("word: " + str + ", occur times: " + Collections.frequency(word_list, str));
 88 //            }
 89 
 90             // HashMap
 91             Map<String, Integer> hashMap = new HashMap<String, Integer>();
 92             for (String temp : word_list) {
 93                 Integer count = hashMap.get(temp);
 94                 hashMap.put(temp, (count == null) ? 1 : count + 1);
 95             }
 96 
 97             // TreeMap
 98             TreeMap<String, Integer> treeMap = new TreeMap<String, Integer>(hashMap);
 99 
100             // Record result to another file
101             printMap(treeMap);
102         } catch (IOException e) {
103             e.printStackTrace();
104         } finally {
105             closeInputStream(br);
106             closeOutputStream(bw);
107         }
108     }
109 
110     //将数据写入文件（两种打印方式）
111     public static void printMap(Map<String, Integer> map) throws IOException {
112 
113         bw = new BufferedWriter(new FileWriter(path_result));
114 
115         Set<String> keys = map.keySet();
116         for (String s : keys) {
117             System.out.println("word: " + s + ", times: " + map.get(s));
118             writeResult("word: " + s + ", times: " + map.get(s));
119         }
120         System.out.println();
121         System.out.println("单词数目：" + countWord);
122         writeResult(String.valueOf(countWord));
123         System.out.println("字符数目：" + countChar);
124         writeResult(String.valueOf(countChar));
125         System.out.println("行数：" + countLine);
126         writeResult(String.valueOf(countLine));
127 //        for (Map.Entry<String, Integer> entry : map.entrySet()) {
128 //            System.out.println("word: " + entry.getKey() + ", number : " + entry.getValue());
129 //            writeResult("word: " + entry.getKey() + ", number : " + entry.getValue());
130 //        }
131 
132     }
133 
134     //写入函数
135     public static void writeResult(String line) throws IOException {
136 
137         try {
138             if (bw != null) {
139                 bw.write(line);
140                 bw.newLine();
141                 bw.flush();
142             }
143         } catch (IOException e) {
144             e.printStackTrace();
145             closeOutputStream(bw);
146         }
147     }
148 
149     public static void closeOutputStream(BufferedWriter writer) {
150         try {
151             if (writer != null) {
152                 writer.close();
153             }
154         } catch (IOException e) {
155             e.printStackTrace();
156         }
157     }
158 
159     public static void closeInputStream(BufferedReader reader) {
160         try {
161             if (reader != null) {
162                 reader.close();
163             }
164         } catch (IOException e) {
165             e.printStackTrace();
166         }
167     }
168 }

经过修改的代码在原有功能的基础上，添加了能够统计单词数目，字符数目，文件行数的功能。优化输出打印方式，使结果查看更简单。