下面是测试的代码:
public class ArrayCopyPerformanceTest {
public static class Person {
public String name;
}
public static void main(String[] args) {
int arraySize = 1000;
int testTimes = 100000;
Person[] ps1 = genPersons(arraySize);
long t1 = System.currentTimeMillis();
Person[] ps2 = new Person[arraySize];
for (int i = 0; i < testTimes; i++) {
System.arraycopy(ps1, 0, ps2, 0, arraySize);
}
long t2 = System.currentTimeMillis();
System.out.println("time used by System.arraycopy: " + (t2 - t1) + "ms");
long t = System.currentTimeMillis();
Person[] ps = new Person[arraySize];
for (int i = 0; i < testTimes; i++) {
ps2 = ps1.clone();
}
long tt = System.currentTimeMillis();
System.out.println("time used by clone: " + (tt - t) + "ms");
t1 = System.currentTimeMillis();
ps2 = new Person[arraySize];
for (int i = 0; i < testTimes; i++) {
for (int j = 0; j < arraySize; j++) {
ps2[j] = ps1[j];
}
}
t2 = System.currentTimeMillis();
System.out.println("time used by manual loop: " + (t2 - t1) + "ms");
}
public static Person[] genPersons(int arraySize) {
Person[] ps = new Person[arraySize];
for (int i = 0; i < arraySize; i++) {
Person p = new Person();
p.name = "james" + i;
ps[i] = p;
}
return ps;
}
}
执行的结果:
time used by System.arraycopy: 116ms
time used by clone: 370ms
time used by manual loop: 641ms
从结果来看性能还是会有不小的差距的,clone虽然在性能上较之普通循环赋值有所提升,但是,不够灵活,不能指定赋值的元素,而且,java内置的集合类型,eg,ArrayList,内部的add()和remove()方法的内部实现都是用arraycopy的。所以,综合来说,对于批量数据的赋值推荐用arraycopy。