caffeine的基本使用

最新推荐文章于 2025-06-27 17:30:51 发布

转载最新推荐文章于 2025-06-27 17:30:51 发布 · 2.3k 阅读

3 ·

CC 4.0 BY-SA版权

原文链接：https://www.kancloud.cn/ichenpeng/blog/1094838

文章标签：

#java

本地缓存专栏收录该内容

1 篇文章

订阅专栏

Caffeine是一款基于JDK1.8的高性能缓存库，提供过期策略和装载策略。常见用法包括设置过期时间和最大容量，以及使用removalListener监听清理事件。填充策略包括手动、自动、异步手动和异步自动加载。驱逐策略可以基于容量大小或时间，例如写入后10秒过期或写入/访问后无操作10秒过期。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

java内存类库caffeine

caffeine是基于jdk1.8上开发的高性能缓存管理类库(程序内存)，它和 ConcurrentMap类似，但是增加了过期策略、装载策略等独有的东西。

此文章转载自：https://www.kancloud.cn/ichenpeng/blog/1094838

1. 常见用法

public static void main(String... args) throws Exception {
Cache<String, String> cache = Caffeine.newBuilder()
        //5秒没有读写自动删除
        .expireAfterAccess(5, TimeUnit.SECONDS)
        //最大容量1024个，超过会自动清理空间
        .maximumSize(1024)
        .removalListener(((key, value, cause) -> {
            //清理通知 key,value ==> 键值对   cause ==> 清理原因
        }))
        .build();

//添加值
cache.put("张三", "浙江");
//获取值
cache.getIfPresent("张三");
//remove
cache.invalidate("张三");
}

2. 填充策略

填充策略是指如何在key不存在的情况下，如何创建一个对象进行返回，主要分为下面四种

2.1 手动(Manual)

public static void main(String... args) throws Exception {
        Cache<String, Integer> cache = Caffeine.newBuilder().build();

        Integer age1 = cache.getIfPresent("张三");
        System.out.println(age1);

        //当key不存在时，会立即创建出对象来返回，age2不会为空
        Integer age2 = cache.get("张三", k -> {
            System.out.println("k:" + k);
            return 18;
        });
        System.out.println(age2);
}

null
k:张三
18

2.2 自动(Loading)

public static void main(String... args) throws Exception {

    //此时的类型是 LoadingCache 不是 Cache
    LoadingCache<String, Integer> cache = Caffeine.newBuilder().build(key -> {
        System.out.println("自动填充:" + key);
        return 18;
    });

    Integer age1 = cache.getIfPresent("张三");
    System.out.println(age1);

    // key 不存在时 会根据给定的CacheLoader自动装载进去
    Integer age2 = cache.get("张三");
    System.out.println(age2);
}

null
自动填充:张三
18

2.3 异步手动(Asynchronous Manual)

public static void main(String... args) throws Exception {
    AsyncCache<String, Integer> cache = Caffeine.newBuilder().buildAsync();

    //会返回一个 future对象， 调用future对象的get方法会一直卡住直到得到返回，和多线程的submit一样
    CompletableFuture<Integer> ageFuture = cache.get("张三", name -> {
        System.out.println("name:" + name);
        return 18;
    });

    Integer age = ageFuture.get();
    System.out.println("age:" + age);
}

name:张三
age:18

2.4 异步自动(Asynchronously Loading)

public static void main(String... args) throws Exception {
    //和1.4基本差不多
    AsyncLoadingCache<String, Integer> cache = Caffeine.newBuilder().buildAsync(name -> {
        System.out.println("name:" + name);
        return 18;
    });
    CompletableFuture<Integer> ageFuture = cache.get("张三");

    Integer age = ageFuture.get();
    System.out.println("age:" + age);
}

name:张三
age:18

3. 驱逐策略

3.1 基于容量大小

public static void main(String... args) throws Exception {
    Cache<String, Integer> cache = Caffeine.newBuilder().maximumSize(10)
    .removalListener((key, value, cause) -> {
        System.out.println(String.format("key %s was removed %s / %s", key, value, cause));
    }).build();

    for (int i = 0; i < 100; i++) {
        cache.put("name" + i, i);
    }
    Thread.currentThread().join();
}

key name0 was removed 0 / SIZE
key name3 was removed 3 / SIZE
key name4 was removed 4 / SIZE
key name5 was removed 5 / SIZE
key name39 was removed 39 / SIZE
key name38 was removed 38 / SIZE
key name37 was removed 37 / SIZE
key name36 was removed 36 / SIZE
key name35 was removed 35 / SIZE
key name34 was removed 34 / SIZE
key name33 was removed 33 / SIZE
key name32 was removed 32 / SIZE
...

3.2 基于时间

//写入后10秒过期,重新写入会刷新过期时间
Cache<String, Integer> cache1 = Caffeine.newBuilder().expireAfterWrite(10, TimeUnit.SECONDS).build();
//写入或读取后10秒无任务操作会过期，读写都会刷新过期时间
Cache<String, Integer> cache2 = Caffeine.newBuilder().expireAfterAccess(10, TimeUnit.SECONDS).build();