一、测试代码
public class XY_ThreadData
{
private static Integer data = 0;
private static Map<Thread, Integer> map = new HashMap<Thread, Integer>();
private static ThreadLocal<Integer> local = new ThreadLocal<Integer>();
public static void setData(Integer value)
{
data = value;
}
public static Integer getData()
{
System.out.println("ThreadName:" + Thread.currentThread() + " data value:" + data);
return data;
}
public static void setMapData(Integer value)
{
map.put(Thread.currentThread(), value);
}
public static Integer getMapData()
{
Object obj = map.get(Thread.currentThread());
System.out.println("ThreadName:" + Thread.currentThread() + "map value:" + obj);
return Integer.parseInt(obj.toString());
}
public static void setThreadLocalData(Integer value)
{
local.set(value);
}
public static Integer getThreadLocalData()
{
Object obj = local.get();
System.out.println("ThreadName:" + Thread.currentThread() + "threadlocal value:" + obj);
return Integer.parseInt(obj.toString());
}
}
public class XY_ThreadData_Test
{
public static void main(String[] args)
{
for (int i = 0; i < 10; i++)
{
new Thread(new Runnable() {
public void run()
{
final int value = new Random().nextInt(); // 每个线程自己创建的变量
XY_ThreadData.setData(value);
XY_ThreadData.getData();
}
}).start();
}
}
}
ThreadName:Thread[Thread-3,5,main] data value:1046062244
ThreadName:Thread[Thread-6,5,main] data value:-879673875
ThreadName:Thread[Thread-2,5,main] data value:-125397465
ThreadName:Thread[Thread-4,5,main] data value:-1546413071
ThreadName:Thread[Thread-0,5,main] data value:754770101
ThreadName:Thread[Thread-8,5,main] data value:-1666786926
ThreadName:Thread[Thread-5,5,main] data value:-1666786926
ThreadName:Thread[Thread-9,5,main] data value:1046062244
ThreadName:Thread[Thread-1,5,main] data value:269410746
ThreadName:Thread[Thread-7,5,main] data value:269410746
分析:可以看到以下两个线程中data的值时一样的,没有做到各个线程变量独一份
ThreadName:Thread[Thread-8,5,main] data value:-1666786926
ThreadName:Thread[Thread-5,5,main] data value:-1666786926
public class XY_ThreadData_Test
{
public static void main(String[] args)
{
for (int i = 0; i < 10; i++)
{
new Thread(new Runnable() {
public void run()
{
final int value = new Random().nextInt();
XY_ThreadData.setMapData(value);
XY_ThreadData.getMapData();
}
}).start();
}
}
}
ThreadName:Thread[Thread-0,5,main]map value:-1138167111
ThreadName:Thread[Thread-4,5,main]map value:-1545929782
ThreadName:Thread[Thread-6,5,main]map value:-1612385717
ThreadName:Thread[Thread-3,5,main]map value:-1390594683
ThreadName:Thread[Thread-8,5,main]map value:518506934
ThreadName:Thread[Thread-2,5,main]map value:1583239372
ThreadName:Thread[Thread-5,5,main]map value:995578601
ThreadName:Thread[Thread-1,5,main]map value:-916627474
ThreadName:Thread[Thread-7,5,main]map value:-960206804
ThreadName:Thread[Thread-9,5,main]map value:-1187504747
分析:模拟线程变量独一份,无重复
public class XY_ThreadData_Test
{
public static void main(String[] args)
{
for (int i = 0; i < 10; i++)
{
new Thread(new Runnable() {
public void run()
{
final int value = new Random().nextInt();
XY_ThreadData.setThreadLocalData(value);
XY_ThreadData.getThreadLocalData();
}
}).start();
}
}
}
ThreadName:Thread[Thread-1,5,main]threadlocal value:935024745
ThreadName:Thread[Thread-4,5,main]threadlocal value:1207176846
ThreadName:Thread[Thread-7,5,main]threadlocal value:-1503260374
ThreadName:Thread[Thread-9,5,main]threadlocal value:-1538563684
ThreadName:Thread[Thread-6,5,main]threadlocal value:955259906
ThreadName:Thread[Thread-8,5,main]threadlocal value:894428541
ThreadName:Thread[Thread-3,5,main]threadlocal value:730986356
ThreadName:Thread[Thread-2,5,main]threadlocal value:-540225655
ThreadName:Thread[Thread-5,5,main]threadlocal value:-2003809947
ThreadName:Thread[Thread-0,5,main]threadlocal value:1917431015
分析:线程变量独一份,无重复
二、ThreadLocal分析
要点1
ThreadLocal不是用来解决共享对象的多线程访问问题的,一般情况下通过ThreadLocal.set()到线程中的对象是该线程自己使用的对象,其他线程是不需要访问的,也访问不到的。各个线程中访问的是不同的对象。
要点2
说ThreadLocal使得各线程能够保持各自独立的一个对象,并不是通过ThreadLocal.set()来实现的,而是通过每个线程中的new对象的操作来创建的对象,每个线程创建一个,不是什么对象的拷贝或副本。通ThreadLocal.set()将这个新创建的对象的引用保存到各线程的自己的一个map中,每个线程都有这样一个map,执行ThreadLocal.get()时,各线程从自己的map中取出放进去的对象,因此取出来的是各自自己线程中的对象,ThreadLocal实例是作为map的key来使用的。
要点3
如果ThreadLocal.set()进去的东西本来就是多个线程共享的同一个对象,那么多个线程的ThreadLocal.get()取得的还是这个共享对象本身,还是有并发访问问题。
更多关于ThreadLocal信息请参看:http://www.iteye.com/topic/103804