java线程之ThreadLocal

最新推荐文章于 2022-05-08 10:14:44 发布

原创最新推荐文章于 2022-05-08 10:14:44 发布 · 654 阅读

0 ·

CC 4.0 BY-SA版权

java 专栏收录该内容

121 篇文章

订阅专栏

本文深入探讨了ThreadLocal的概念及其在多线程编程中的使用方法，通过实例展示了如何确保每个线程拥有独立的共享资源，从而实现线程间的隔离与互不影响。重点介绍了ThreadLocal如何在不改变全局变量的情况下，为每个线程提供专属的资源，适用于需要维护线程间数据独立性的场景。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

ThreadLocal是指某线程的局部变量。用法如下：

package com.concurrent;

import junit.framework.TestCase;

public class TestThreadLocal extends TestCase {

	public void testThreadLocal1() {
		SharedResource sharedResource = new SharedResource();
		for (int i = 0; i < 3; i++) {
			TestThread t = new TestThread(sharedResource);
			t.start();
		}
	}
	
	public void testThreadLocal2() {
		for (int i = 0; i < 3; i++) {
			SharedResource sharedResource = new SharedResource();
			TestThread t = new TestThread(sharedResource);
			t.start();
		}
	}
	
	public void testThreadLocal3() {
		SharedResource sharedResource = new ThreadLocalSharedResource();
		for (int i = 0; i < 3; i++) {
			TestThread t = new TestThread(sharedResource);
			t.start();
		}
	}

	static class SharedResource {
		private int count = 0;

		public int getCount() {
			return count++;
		}
	}
	
	static class ThreadLocalSharedResource extends SharedResource {
		private ThreadLocal<Integer> threadLocal = new ThreadLocal<Integer>() {

			@Override
			protected Integer initialValue() {
				return 0;
			}
			
		};

		@Override
		public int getCount() {
			int oldValue = threadLocal.get();
			int newValue = oldValue + 1;
			threadLocal.set(newValue);
			return oldValue;
		}
		
		
	}

	static class TestThread extends Thread {

		private SharedResource sharedResource;

		TestThread(SharedResource sharedResource) {
			this.sharedResource = sharedResource;
		}

		@Override
		public void run() {
			for (int i = 0; i < 3; i++) {
				System.out.println(Thread.currentThread() + ", " + sharedResource.getCount());
			}
		}

	}
}

运行testThreadLocal1，输出：

Thread[Thread-0,5,main], 0
Thread[Thread-0,5,main], 1
Thread[Thread-0,5,main], 2
Thread[Thread-1,5,main], 3
Thread[Thread-1,5,main], 4
Thread[Thread-1,5,main], 5
Thread[Thread-2,5,main], 6
Thread[Thread-2,5,main], 7
Thread[Thread-2,5,main], 8

一个共享资源，三个线程共享。

如果要的结果是每个线程只会输出0,1,2，那可以new三个共享资源，也就是testThreadLocal2，输出结果：

Thread[Thread-0,5,main], 0
Thread[Thread-0,5,main], 1
Thread[Thread-0,5,main], 2
Thread[Thread-2,5,main], 0
Thread[Thread-2,5,main], 1
Thread[Thread-2,5,main], 2
Thread[Thread-1,5,main], 0
Thread[Thread-1,5,main], 1
Thread[Thread-1,5,main], 2

但是，共享资源一般只有一个，那怎么保证上面的输出呢？那就要用到ThreadLocal了，运行testThreadLocal3：

Thread[Thread-0,5,main], 0
Thread[Thread-0,5,main], 1
Thread[Thread-0,5,main], 2
Thread[Thread-2,5,main], 0
Thread[Thread-2,5,main], 1
Thread[Thread-2,5,main], 2
Thread[Thread-1,5,main], 0
Thread[Thread-1,5,main], 1
Thread[Thread-1,5,main], 2

ThreadLocal保证每个线程里的共享资源都是独立的。