HashMap的clear()方法和new HashMap的效率问题

最近研究Lucene的时候,遇到的用到大量Map的问题,心生好奇,想看一下在1W,10W,100W三种数据量下,new HashMap ,与 HashMap.clear()方法的效率问题。

提前说明:就是自己好奇测着玩的,不能作为任何依据,只是提供一个参考。

JDK版本:1.8

IDE:idea

 

Clear()

数据量级:1W ,执行次数:10次,取10次时间的平均值

执行结果:

map-clear:27
map-clear:27
map-clear:15
map-clear:14
map-clear:15
map-clear:13
map-clear:11
map-clear:3
map-clear:4
map-clear:4
平均时间:13

执行内存快照:

 

数据量级:10W ,执行次数:10次,取10次时间的平均值

执行结果:

map-clear:212
map-clear:128
map-clear:99
map-clear:161
map-clear:49
map-clear:58
map-clear:44
map-clear:112
map-clear:218
map-clear:40
平均时间:112

执行内存快照:

 

数据量级:100W ,执行次数:10次,取10次时间的平均值

执行结果:

map-clear:661
map-clear:657
map-clear:364
map-clear:426
map-clear:343
map-clear:481
map-clear:271
map-clear:212
map-clear:176
map-clear:165
平均时间:375

执行内存快照:

new HashMap()

数据量级:1W ,执行次数:10次,取10次时间的平均值

执行结果:

map-new:27
map-new:20
map-new:10
map-new:13
map-new:19
map-new:21
map-new:9
map-new:7
map-new:7
map-new:6
平均时间:13

执行内存快照:

数据量级:10W ,执行次数:10次,取10次时间的平均值

执行结果:

map-new:122
map-new:58
map-new:66
map-new:110
map-new:83
map-new:45
map-new:62
map-new:36
map-new:29
map-new:30
平均时间:64

执行内存快照:

数据量级:100W ,执行次数:10次,取10次时间的平均值

执行结果:

map-new:539
map-new:493
map-new:420
map-new:294
map-new:455
map-new:305
map-new:379
map-new:256
map-new:314
map-new:315
平均时间:377

执行内存快照:

 

测试类:

public class MapTest {


    public static void main(String[] args) throws InterruptedException {
        
        long totalTimes = 0L;
        for (int i = 0 ; i < 10 ; i++){
            totalTimes += mapToNew(1000000L);
        }
        System.out.println("平均时间:" + (totalTimes / 10));
        Thread.sleep(100000L);
    }


    /**
     * new map
     * @param size
     * @return
     */
    public static long mapToNew(long size)  {
        long nowTime = new Date().getTime();
        Map<String,Object> map;
        for (int i = 0 ; i < size ; i++){
            map = new HashMap<>();
            map.put("id",i);
            map.put("name","测试文章");
            map.put("realname","测试文章11111");
            map.put("text","测试文章+++++++++++++++++++");
            map.put("long",4324L);
            map.put("double",4354D);
            map.put("text2","测试文章+++++++++++++++++++");
            map.put("text3","测试文章+++++++++++++++++++");
            map.put("text4","测试文章+++++++++++++++++++");
            map.put("text5","测试文章+++++++++++++++++++");
            map.put("text6","测试文章+++++++++++++++++++");
            map.put("text7","测试文章+++++++++++++++++++");
        }
        long timeSize = new Date().getTime() - nowTime;
        System.out.println("map-new:" + timeSize);
        return timeSize;
    }


    /**
     * clear map
     * @param size
     * @return
     */
    public static long mapToClear(long size)  {
        long nowTime = new Date().getTime();
        Map<String,Object> map = new HashMap<>();
        for (int i = 0 ; i < size ; i++){
            map.clear();
            map.put("id",i);
            map.put("name","测试文章");
            map.put("realname","测试文章11111");
            map.put("text","测试文章+++++++++++++++++++");
            map.put("long",4324L);
            map.put("double",4354D);
            map.put("text2","测试文章+++++++++++++++++++");
            map.put("text3","测试文章+++++++++++++++++++");
            map.put("text4","测试文章+++++++++++++++++++");
            map.put("text5","测试文章+++++++++++++++++++");
            map.put("text6","测试文章+++++++++++++++++++");
            map.put("text7","测试文章+++++++++++++++++++");
        }
        long timeSize = new Date().getTime() - nowTime;
        System.out.println("map-clear:" + timeSize);
        return timeSize;
    }

}

 

posted @ 2018-11-23 15:12  InkYi  阅读(965)  评论(0编辑  收藏  举报