java List遍歷性能比較

一、測試代碼

public class ListPerformanceTest {
    public static void main(String[] args) {
        List<Integer> arrayList = new ArrayList<>();
        List<Integer> linkedList = new LinkedList<>();
        for (int i = 0; i < 1000000; ++i) {
            arrayList.add(i);
            linkedList.add(i);
        }
        test1(arrayList);
        test2(arrayList);
        test3(arrayList);

        test1(linkedList);
        test2(linkedList);
        test3(linkedList);
    }
    //for循環測試
    public static void test1(List<Integer> list) {
        long start = System.currentTimeMillis();
        for (int i = 0; i < list.size(); ++i) {
            int num = list.get(i);
        }
        long end = System.currentTimeMillis();
        System.out.println("test1 : " + (end - start));
    }
    //iterator測試
    public static void test2(List<Integer> list) {
        long start = System.currentTimeMillis();
        Iterator iterator = list.iterator();
        while (iterator.hasNext()) {
            int num = (Integer) iterator.next();
        }
        long end = System.currentTimeMillis();
        System.out.println("test2 : " + (end - start));
    }
    //for each測試
    public static void test3(List<Integer> list) {
        long start = System.currentTimeMillis();
        for (Integer integer : list) {
            int num = integer;
        }
        long end = System.currentTimeMillis();
        System.out.println("test3 : " + (end - start));
    }
}

二、ArrayList遍歷性能比較（時間單位：微秒）

size	10	100	1,000	10,000	100,000	1,000,000
for循環	0	0	1	2	10	13
Iterator	1	0	0	2	11	14
for each	0	0	0	2	11	12

三、LinkedList遍歷性能比較（時間單位：微秒）

size	10	100	1,000	10,000	100,000	1,000,000
for循環	0	0	3	74	8656	>15min
Iterator	1	1	1	2	4	26
for each	0	0	1	3	4	29

四、總結

1. 對於ArrayList，無論size是多大，耗時都差不多，選擇哪種方式都一樣
2. for each內部其實也是用iterator實現的，所以不管是ArrayList還是LinkedList，iterator和for each的性能都差不多
3. 由於LinkedList每次get的時候，都得從開始位置查找，所以遍歷的的性能會很慢
4. 綜合來看，建議用for each，代碼簡潔，性能也不差

五、參考資料

https://blog.csdn.net/dengnanhua/article/details/64692191

https://blog.csdn.net/xyc_csdn/article/details/69683799