在一个数据库设计里，假如使用了逻辑主键，那么你一般都需要一个ID生成器去生成逻辑主键。
在许多数据库里面，都提供了ID生成的机制，如Oracle中的sequence，MSSQL中的identity，可惜这些方法各种数据库都不同的，所以很多人愿意找寻一种通用的方式。
编写代码，1、2、3……这样一直累加是最直接的想法，Java用以下方式去实现
private static AtomicInteger uniqueId = new AtomicInteger(0);
public static String nextId() {
return Integer.toString(uniqueId.incrementAndGet());
}
当然，这样太简单了，并且一重新启动，计数器就归 0 了，一般的做法可以用 时间 + 计数器 的方式，
private static final long ONE_STEP = 10;
private static final long BASE = 1129703383453l;
private static final Lock LOCK = new ReentrantLock();
private static long lastTime = System.currentTimeMillis();
private static short lastCount = 0;
/**
* a time (as returned by {@link System#currentTimeMillis()}) at which
* the VM that this UID was generated in was alive
* @serial
*/
private final long time;
/**
* 16-bit number to distinguish UID instances created
* in the same VM with the same time value
* @serial
*/
private final short count;
/**
* Generates a UID that is unique over time with
* respect to the host that it was generated on.
*/
public UID() {
LOCK.lock();
try {
if (lastCount == ONE_STEP) {
boolean done = false;
while (!done) {
long now = System.currentTimeMillis();
if (now == lastTime) {
// pause for a second to wait for time to change
try {
Thread.currentThread().sleep(1);
}
catch (java.lang.InterruptedException e) {
} // ignore exception
continue;
}
else {
lastTime = now;
lastCount = 0;
done = true;
}
}
}
time = lastTime;
count = lastCount++;
}
finally {
LOCK.unlock();
}
}
在一个群集的环境里面，通常还需要加上IP的前缀，即 IP + 时间 + 计数器，这个就是JAVA原版本的实现了。
但是，觉得这个ID太长了？那没办法，回到用数据库的方法吧。