Java并发编程之ThreadLocal类

ThreadLocal类可以理解为ThreadLocalVariable(线程局部变量),提供了get与set等访问接口或方法,这些方法为每个使用该变量的线程都存有一份独立的副本,因此get总是返回当前执行线程在调用set时设置的最新值。可以将ThreadLocal<T>视为 包含了Map<Thread,T>对象,保存了特定于该线程的值。

概括起来说,对于多线程资源共享的问题,同步机制采用了“以时间换空间”的方式,而ThreadLocal采用了“以空间换时间”的方式。前者仅提供一份变量,让不同的线程排队访问,而后者为每一个线程都提供了一份变量,因此可以同时访问而互不影响。

模拟ThreadLocal

    import java.util.Collections;
    import java.util.HashMap;
    import java.util.Map;
     
    public class SimpleThreadLocal<T> {
    private Map<Thread, T> valueMap = Collections
    .synchronizedMap(new HashMap<Thread, T>());
     
    public void set(T newValue) {
    valueMap.put(Thread.currentThread(), newValue); // ①键为线程对象,值为本线程的变量副本
    }
     
    public T get() {
    Thread currentThread = Thread.currentThread();
    T o = valueMap.get(currentThread); // ②返回本线程对应的变量
    if (o == null && !valueMap.containsKey(currentThread)) { // ③如果在Map中不存在,放到Map中保存起来。
    o = initialValue();
    valueMap.put(currentThread, o);
    }
    return o;
    }
     
    public void remove() {
    valueMap.remove(Thread.currentThread());
    }
     
    protected T initialValue() {
    return null;
    }
    }
实用ThreadLocal
    class Count {
    private SimpleThreadLocal<Integer> count = new SimpleThreadLocal<Integer>() {
    @Override
    protected Integer initialValue() {
    return 0;
    }
    };
     
    public Integer increase() {
    count.set(count.get() + 1);
    return count.get();
    }
     
    }
     
    class TestThread implements Runnable {
    private Count count;
     
    public TestThread(Count count) {
    this.count = count;
    }
     
    @Override
    public void run() {
    // TODO Auto-generated method stub
    for (int i = 1; i <= 3; i++) {
    System.out.println(Thread.currentThread().getName() + "\t" + i
    + "th\t" + count.increase());
    }
    }
    }
     
    public class TestThreadLocal {
    public static void main(String[] args) {
    Count count = new Count();
    Thread t1 = new Thread(new TestThread(count));
    Thread t2 = new Thread(new TestThread(count));
    Thread t3 = new Thread(new TestThread(count));
    Thread t4 = new Thread(new TestThread(count));
    t1.start();
    t2.start();
    t3.start();
    t4.start();
    }
    }

输出
Thread-0    1th    1
Thread-0    2th    2
Thread-0    3th    3
Thread-3    1th    1
Thread-1    1th    1
Thread-1    2th    2
Thread-2    1th    1
Thread-1    3th    3
Thread-3    2th    2
Thread-3    3th    3
Thread-2    2th    2

Thread-2    3th    3  

上一篇:基于TFS实践敏捷-可视化管理


下一篇:Redis乐观锁解决超卖问题