foreach 相当于一个迭代器——它是同一事物的语法糖。所以你应该尽可能选择foreach而不是迭代器,因为它很方便并且代码更简洁。
我在另一个答案中写过这个:https://stackoverflow.com/questions/22110482/uses-and-syntax-for-for-each-loop-in-java/22110517#22110517
正如@RonnyShapiro 所说,在某些情况下您需要使用迭代器,但在许多情况下foreach 就足够了。请注意,foreach不是普通的for 循环。当需要访问索引时,需要正常的for 循环。尽管您可以使用 foreach 手动创建一个单独的索引 int 变量,但从 variable-scope 的角度来看,这并不理想。
这里有更多信息:Which is more efficient, a for-each loop, or an iterator?
访问集合时,foreach 是 significantly faster,而不是基本的 for 循环的数组访问。然而,当访问数组时——至少使用原始数组和包装数组——通过索引进行访问方式更快。见下文。
在访问 int 或 Integer 数组时,索引比迭代器快 23-40%。这是下面测试类的输出,它对 100 个元素的原始整数数组中的数字求和(A 是迭代器,B 是索引):
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 358,597,622 nanoseconds
Test B: 269,167,681 nanoseconds
B faster by 89,429,941 nanoseconds (24.438799231635727% faster)
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 377,461,823 nanoseconds
Test B: 278,694,271 nanoseconds
B faster by 98,767,552 nanoseconds (25.666236154695838% faster)
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 288,953,495 nanoseconds
Test B: 207,050,523 nanoseconds
B faster by 81,902,972 nanoseconds (27.844689860906513% faster)
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 375,373,765 nanoseconds
Test B: 283,813,875 nanoseconds
B faster by 91,559,890 nanoseconds (23.891659337194227% faster)
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 375,790,818 nanoseconds
Test B: 220,770,915 nanoseconds
B faster by 155,019,903 nanoseconds (40.75164734599769% faster)
[C:\java_code\]java TimeIteratorVsIndexIntArray 1000000
Test A: 326,373,762 nanoseconds
Test B: 202,555,566 nanoseconds
B faster by 123,818,196 nanoseconds (37.437545972215744% faster)
完整的测试类:
import java.text.NumberFormat;
import java.util.Locale;
/**
<P>{@code java TimeIteratorVsIndexIntArray 1000000}</P>
@see <CODE><A HREF="https://stackoverflow.com/questions/180158/how-do-i-time-a-methods-execution-in-java">https://stackoverflow.com/questions/180158/how-do-i-time-a-methods-execution-in-java</A></CODE>
**/
public class TimeIteratorVsIndexIntArray {
public static final NumberFormat nf = NumberFormat.getNumberInstance(Locale.US);
public static final void main(String[] tryCount_inParamIdx0) {
int testCount;
//Get try-count from command-line parameter
try {
testCount = Integer.parseInt(tryCount_inParamIdx0[0]);
} catch(ArrayIndexOutOfBoundsException | NumberFormatException x) {
throw new IllegalArgumentException("Missing or invalid command line parameter: The number of testCount for each test. " + x);
}
//Test proper...START
int[] intArray = new int[] {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100};
long lStart = System.nanoTime();
for(int i = 0; i < testCount; i++) {
testIterator(intArray);
}
long lADuration = outputGetNanoDuration("A", lStart);
lStart = System.nanoTime();
for(int i = 0; i < testCount; i++) {
testFor(intArray);
}
long lBDuration = outputGetNanoDuration("B", lStart);
outputGetABTestNanoDifference(lADuration, lBDuration, "A", "B");
}
private static final void testIterator(int[] int_array) {
int total = 0;
for(int i = 0; i < int_array.length; i++) {
total += int_array[i];
}
}
private static final void testFor(int[] int_array) {
int total = 0;
for(int i : int_array) {
total += i;
}
}
//Test proper...END
//Timer testing utilities...START
public static final long outputGetNanoDuration(String s_testName, long l_nanoStart) {
long lDuration = System.nanoTime() - l_nanoStart;
System.out.println("Test " + s_testName + ": " + nf.format(lDuration) + " nanoseconds");
return lDuration;
}
public static final long outputGetABTestNanoDifference(long l_aDuration, long l_bDuration, String s_aTestName, String s_bTestName) {
long lDiff = -1;
double dPct = -1.0;
String sFaster = null;
if(l_aDuration > l_bDuration) {
lDiff = l_aDuration - l_bDuration;
dPct = 100.00 - (l_bDuration * 100.0 / l_aDuration + 0.5);
sFaster = "B";
} else {
lDiff = l_bDuration - l_aDuration;
dPct = 100.00 - (l_aDuration * 100.0 / l_bDuration + 0.5);
sFaster = "A";
}
System.out.println(sFaster + " faster by " + nf.format(lDiff) + " nanoseconds (" + dPct + "% faster)");
return lDiff;
}
//Timer testing utilities...END
}
我还为 Integer 数组运行了这个,索引仍然是明显的赢家,但速度只提高了 18% 到 25%。
但是,对于 Integers 中的 List,迭代器更快。只需将上面代码中的int-array改为
List<Integer> intList = Arrays.asList(new Integer[] {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100});
并对测试功能进行必要的更改(int[] 到 List<Integer>、length 到 size() 等)
[C:\java_code\]java TimeIteratorVsIndexIntegerList 1000000
Test A: 3,429,929,976 nanoseconds
Test B: 5,262,782,488 nanoseconds
A faster by 1,832,852,512 nanoseconds (34.326681820485675% faster)
[C:\java_code\]java TimeIteratorVsIndexIntegerList 1000000
Test A: 2,907,391,427 nanoseconds
Test B: 3,957,718,459 nanoseconds
A faster by 1,050,327,032 nanoseconds (26.038700083921256% faster)
[C:\java_code\]java TimeIteratorVsIndexIntegerList 1000000
Test A: 2,566,004,688 nanoseconds
Test B: 4,221,746,521 nanoseconds
A faster by 1,655,741,833 nanoseconds (38.71935684115413% faster)
[C:\java_code\]java TimeIteratorVsIndexIntegerList 1000000
Test A: 2,770,945,276 nanoseconds
Test B: 3,829,077,158 nanoseconds
A faster by 1,058,131,882 nanoseconds (27.134122749113843% faster)
[C:\java_code\]java TimeIteratorVsIndexIntegerList 1000000
Test A: 3,467,474,055 nanoseconds
Test B: 5,183,149,104 nanoseconds
A faster by 1,715,675,049 nanoseconds (32.60101667104192% faster)
[C:\java_code\]java TimeIteratorVsIndexIntList 1000000
Test A: 3,439,983,933 nanoseconds
Test B: 3,509,530,312 nanoseconds
A faster by 69,546,379 nanoseconds (1.4816434912159906% faster)
[C:\java_code\]java TimeIteratorVsIndexIntList 1000000
Test A: 3,451,101,466 nanoseconds
Test B: 5,057,979,210 nanoseconds
A faster by 1,606,877,744 nanoseconds (31.269164666060377% faster)
在一项测试中,它们几乎是等价的,但迭代器仍然获胜。