多线程与单线程比较

àì夳堔傛蜴生んèń 2024-04-01 18:59 12阅读 0赞

给一个多行的txt文本文件，统计里面每个单词出现多少次（如下图）

![52db7d2ec63e411998b5b08344fb1e71.png][]

**单线程代码：**

import java.io.BufferedReader;
    import java.io.FileNotFoundException;
    import java.io.FileReader;
    import java.io.IOException;
    import java.util.TreeMap;
    
    public class TestDemo {
        public static void main(String[] args) {
    
            FileReader fr = null;
            BufferedReader br = null;
    
            TreeMap<String, Integer> worldMap = new TreeMap<String, Integer>();
            long start = System.currentTimeMillis();
            try {
                fr = new FileReader("fos.txt");
                br = new BufferedReader(fr);
                String temp = "";
                int num = 0;//读取行数
                while ((temp = br.readLine()) != null) {
                    num++;
                    String[] splitStr = temp.trim().split(" ");
                    for (int i = 0; i < splitStr.length; i++) {
                        if (worldMap.containsKey(splitStr[i])) {//原Map中包含key
                            Integer count = worldMap.get(splitStr[i]) + 1;
                            worldMap.put(splitStr[i], count);
                        } else {
                            //原map中不包含这个key（单词）
                            worldMap.put(splitStr[i], 1);
    //                        System.out.println("不包含");
                        }
                    }
                }
                System.out.println("读取文件总行数：" + num);
    
    
                for (String key : worldMap.keySet()) {
                    Integer wordNum = worldMap.get(key);
                    System.out.println(key + "出现的次数为：" + wordNum);
                }
                long totalTime = System.currentTimeMillis() - start;
                System.out.println("程序运行时间：" + totalTime);
    
            } catch (FileNotFoundException e) {
                e.printStackTrace();
            } catch (IOException e) {
                e.printStackTrace();
            } finally {
                try {
                    br.close();
                    fr.close();
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
        }
    }

**运行结果：**

![1a86be64a64c48358a1c91b971dc2121.png][]

**多线程代码：**

import java.util.HashMap;
    import java.util.Map;
    
    public class wordCount implements Runnable {
    
        String content = "";
        Map<String, Integer> map = new HashMap<>();
    
        public wordCount(String content) {
            this.content = content;
        }
    
        @Override
        public void run() {
            String[] split = content.trim().split(" ");
            for (int i = 0; i < split.length; i++) {
                if (map.containsKey(split[i])){
                    Integer value = map.get(split[i]) + 1;
                    map.put(split[i],value);
                }else {
                    map.put(split[i],1);
                }
            }
        }
    }

import java.io.BufferedReader;
    import java.io.FileNotFoundException;
    import java.io.FileReader;
    import java.io.IOException;
    import java.util.HashMap;
    import java.util.Iterator;
    import java.util.Map;
    import java.util.Set;
    import java.util.concurrent.ExecutorService;
    import java.util.concurrent.Executors;
    
    public class wordCountTest {
        static Map<String, Map<String, Integer>> threadMAP = new HashMap<>();
    
        public static void main(String[] args) {
            long start = System.currentTimeMillis();
            FileReader fr = null;
            BufferedReader bf = null;
            int lineNum = 0;
    
            ExecutorService executorService = Executors.newCachedThreadPool();
    
    
            try {
                fr = new FileReader("fos.txt");
                bf = new BufferedReader(fr);
    
                StringBuffer listStrBuf = new StringBuffer("");
                String lineStr = "";
    //            boolean tag = true;
                while ((lineStr = bf.readLine()) != null) {
                    listStrBuf.append(lineStr + " ");
                    lineNum++;
                    if (lineNum % 100000 == 0) {
                        wordCount wordCount = new wordCount(listStrBuf.toString());
                        listStrBuf.delete(0, listStrBuf.length());
                        Thread thread = new Thread(wordCount);
    //                    thread.start();
                        executorService.execute(thread);
                        threadMAP.put("thread-" + lineNum / 100000, wordCount.map);
                    }
                }
    
                if (listStrBuf.toString().length() > 0) {
                    wordCount wordCount = new wordCount(listStrBuf.toString());
                    Thread thread = new Thread(wordCount);
                    executorService.execute(thread);
                    threadMAP.put("thread-last", wordCount.map);
                }
    
                System.out.println("文件总行数为：" + lineNum);
                executorService.shutdown();
    
    
                while (true) {
                    if (executorService.isTerminated()) {
                        Map<String, Integer> mapResult = new HashMap<>();
    //                    for (Map<String, Integer> tempMap : threadMAP.values()) {
    //                        System.out.println(tempMap.keySet());
    //                    }
    
                        for (String key :           //thread-1  thread-2  .......  thread-last
                                threadMAP.keySet()) {
                            Map<String, Integer> perThreadWorldNum = threadMAP.get(key);
                            Set<String> perThreadWords = perThreadWorldNum.keySet();
                            Iterator<String> iterator = perThreadWords.iterator();
                            while (iterator.hasNext()) {
                                String word = iterator.next();
                                if (mapResult.containsKey(word)) {
                                    Integer wordsNum = perThreadWorldNum.get(word);
                                    Integer haveNum = mapResult.get(word);
                                    mapResult.put(word, wordsNum + haveNum);
                                } else {
                                    mapResult.put(word, perThreadWorldNum.get(word));
                                }
                            }
    
                            /*
                            System.out.println(key);
                            System.out.println(threadMAP.get(key).keySet());
                            for (String word :
                                    threadMAP.get(key).keySet()) {
                                System.out.print(word + ":" + threadMAP.get(key).get(word) + " ");
                            }
                            System.out.println();*/
                        }
                        Set<String> resultWords = mapResult.keySet();
                        for (String word : resultWords) {
                            System.out.println(word + " : " + mapResult.get(word));
                        }
    
                        break;
                    }
                }
            } catch (FileNotFoundException e) {
                e.printStackTrace();
            } catch (IOException e) {
                e.printStackTrace();
            } finally {
                try {
                    fr.close();
                    bf.close();
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
    
            long endTime = System.currentTimeMillis() - start;
            System.out.println("程序运行时长：" + endTime + "ms");
        }

运行结果：

![51bea5c50bcb45dab880dd756245e152.png][]

**总结：**

两次运行结果比较：TXT文件总行数33082行，单线程运行时间39ms，多线程运行时间45ms

后面又测试了一个595476行的txt文件，单线程运行时间208ms，多线程运行时间182ms

所以说，并不是多线程运行的时间一定会比单线程运行的时间短。

但是，当文件比较大的时候多线程的优势就体现出来了。

**单线程和多线程在选用的时候要看具体需求。**

[52db7d2ec63e411998b5b08344fb1e71.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/01/72164e69f0c8460fa73b687ce6aea0b8.png
[1a86be64a64c48358a1c91b971dc2121.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/01/703a83dc2a30473aa5762fdc9ea52088.png
[51bea5c50bcb45dab880dd756245e152.png]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/01/62c62c8091e1452d904b6ae66c76ad3e.png