本文章主要体现2点:
1 主线程等待子线程执行完毕。通过CountDownLatch实现
2 演示共享的HashMap的put操作在多个线程中,如何同步。
这里主要是对比了synchronized块和Collections类的装饰功能。
======================================
如下面代码,主线程main,等待现场Thread1和2执行完毕,再打印数据。
注意:由于HashMap本身不是线程安全的,执行结果有时候会很奇怪,如Map的size是2,但只有一个数据。
异常结果如下:
thread already start, waiting...
map size is :2
map key is :2
解决办法是:在MapOper类中,对put进行类同步。
synchronized(this.getClass()){
}
注意,我注释了下面代码,因为他只能同步map引用的操作,而map传入到MapOper后,
map的地址被赋值给this.map了。所以下面的代码不生效。
Collections.synchronizedMap(map);
package multithread;
import java.util.Collections;
import java.util.HashMap;
import java.util.Iterator;
import java.util.concurrent.CountDownLatch;
public class CountDownLatchForWaitSubThread {
@SuppressWarnings("unchecked")
public static void main(String[] args) throws InterruptedException {
CountDownLatchForWaitSubThread test = new CountDownLatchForWaitSubThread();
CountDownLatch latch = new CountDownLatch(2);
HashMap map = new HashMap();
//Collections.synchronizedMap(map);
Thread t1 = new Thread(test.new MapOper(latch,map,1));
Thread t2 = new Thread(test.new MapOper(latch,map,2));
t1.start();
t2.start();
System.out.println("thread already start, waiting...");
latch.await();
for (Iterator iterator = map.keySet().iterator(); iterator.hasNext();) {
String key = (String) iterator.next();
System.out.println("map size is :"+map.size());
System.out.println("map key is :"+key);
}
}
public class MapOper implements Runnable {
CountDownLatch latch ;
HashMap map;
int i;
public MapOper(CountDownLatch latch, HashMap map, int i) {
this.latch = latch;
this.map = map;
this.i = i;
}
public void run() {
synchronized (this.getClass()) {
map.put(String.valueOf(i),i);
}
try {
Thread.sleep(1);
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
latch.countDown();
}
}
}