Java练习：HashMap键值的更新方法对比

在词频统计中，HashMap键值的更新方法有很多种。让我们比较一下运行速度。

 

import java.util.HashMap;

public class Main {
    public static void main(String[] args) {
        String s = "I like JAVA Java java " + "You like JAVA Java java";

        System.out.println("方法1耗时：" + WordCount.method1(s));
        System.out.println("方法2耗时：" + WordCount.method2(s));
        System.out.println("方法3耗时：" + WordCount.method3(s));
        System.out.println("方法4耗时：" + WordCount.method4(s));
        System.out.println("方法5耗时：" + WordCount.method5(s));
    }
}

class WordCount {

    public static long method1(String s) {
        String[] words = s.split(" ");
        HashMap<String, Integer> mapCounter = new HashMap<>();
        long t0 = System.nanoTime();
        for(String word:words) {
            if (mapCounter.containsKey(word)) {
                // 原词频加1
                mapCounter.put(word, mapCounter.get(word) + 1);
            } else {
                //没有这个词，词频1；
                mapCounter.put(word, 1);
            }
        }
        long t1 = System.nanoTime();
        return t1-t0;
    }

    public static long method2(String s) {
        String[] words = s.split(" ");
        HashMap<String, Integer> mapCounter = new HashMap<>();
        long t0 = System.nanoTime();
        for(String word:words) {
            //没有这个词，原词频0；
            mapCounter.putIfAbsent(word, 0);
            // 原词频加1
            mapCounter.put(word, mapCounter.get(word) + 1);  //较快
        }
        long t1 = System.nanoTime();
        return t1-t0;
    }

    public static long method3(String s) {
        String[] words = s.split(" ");
        HashMap<String, Integer> mapCounter = new HashMap<>();
        long t0 = System.nanoTime();
        for(String word:words) {
            //获取原词频，没有这个词，原词频0；原词频加1
            mapCounter.put(word, mapCounter.getOrDefault(word, 0) + 1);  //最快
        }
        long t1 = System.nanoTime();
        return t1-t0;
    }

    public static long method4(String s) {
        String[] words = s.split(" ");
        HashMap<String, Integer> mapCounter = new HashMap<>();
        long t0 = System.nanoTime();
        for(String word:words) {
            //没有这个词，放入这个词，词频1；有这个词，则原词频加1
            mapCounter.merge(word, 1, Integer::sum); //类的静态方法引用，最慢
        }
        long t1 = System.nanoTime();
        return t1-t0;
    }

    public static long method5(String s) {
        String[] words = s.split(" ");
        HashMap<String, Integer> mapCounter = new HashMap<>();
        long t0 = System.nanoTime();
        for(String word:words) {
            //没有这个词，放入这个词，词频1；有这个词，则原词频加1
            mapCounter.merge(word, 1, (a,b)->a+b);  //慢
        }
        long t1 = System.nanoTime();
        return t1-t0;
    }
}

 

 

运行结果：

方法1耗时：36300
方法2耗时：15000
方法3耗时：12300
方法4耗时：57522400
方法5耗时：375000

最快的和最慢的速度相差几千倍。