How does the Java 'for each' loop work?

for (Iterator<String> i = someIterable.iterator(); i.hasNext();) {
    String item = i.next();
    System.out.println(item);
}

Note that if you need to use i.remove(); in your loop, or access the actual iterator in some way, you cannot use the for ( : ) idiom, since the actual iterator is merely inferred.

As was noted by Denis Bueno, this code works for any object that implements the Iterable interface.

Also, if the right-hand side of the for (:) idiom is an array rather than an Iterable object, the internal code uses an int index counter and checks against array.length instead. See the Java Language Specification.

The construct for each is also valid for arrays. e.g.

String[] fruits = new String[] { "Orange", "Apple", "Pear", "Strawberry" };

for (String fruit : fruits) {
    // fruit is an element of the `fruits` array.
}

which is essentially equivalent of

for (int i = 0; i < fruits.length; i++) {
    String fruit = fruits[i];
    // fruit is an element of the `fruits` array.
}

So, overall summary:
[nsayer] The following is the longer form of what is happening:

for(Iterator<String> i = someList.iterator(); i.hasNext(); ) {
  String item = i.next();
  System.out.println(item);
}
Note that if you need to use i.remove(); in your loop, or access the actual iterator in some way, you cannot use the for( : ) idiom, since the actual Iterator is merely inferred.

[Denis Bueno]

It's implied by nsayer's answer, but it's worth noting that the OP's for(..) syntax will work when "someList" is anything that implements java.lang.Iterable -- it doesn't have to be a list, or some collection from java.util. Even your own types, therefore, can be used with this syntax.

The for-each loop, added in Java 5 (also called the "enhanced for loop"), is equivalent to using a java.util.Iterator--it's syntactic sugar for the same thing. Therefore, when reading each element, one by one and in order, a for-each should always be chosen over an iterator, as it is more convenient and concise.

For-each

for (int i : intList) {
   System.out.println("An element in the list: " + i);
}

Iterator

Iterator<Integer> intItr = intList.iterator();
while (intItr.hasNext()) {
   System.out.println("An element in the list: " + intItr.next());
}

There are situations where you must use an Iterator directly. For example, attempting to delete an element while using a for-each can (will?) result in a ConcurrentModificationException.

For-each vs. for-loop: Basic differences

The only practical difference between for-loop and for-each is that, in the case of indexable objects, you do not have access to the index. An example when the basic for-loop is required:

for (int i = 0; i < array.length; i++) {
   if(i < 5) {
      // Do something special
   }  else {
      // Do other stuff
   }
}

Although you could manually create a separate index int-variable with for-each,

int idx = -1;
for (int i : intArray) {
   idx++;
   ...
}

...it is not recommended, since variable-scope is not ideal, and the basic for loop is simply the standard and expected format for this use case.

For-each vs. for-loop: Performance

When accessing collections, a for-each is significantly faster than the basic for loop's array access. When accessing arrays, however--at least with primitive and wrapper-arrays--access via indexes is dramatically faster.

Timing the difference between iterator and index access for primitive int-arrays

Indexes are 23-40 percent faster than iterators when accessing int or Integer arrays. Here is the output from the testing class at the bottom of this post, which sums the numbers in a 100-element primitive-int array (A is iterator, B is index):

[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)

I also ran this for an Integer array, and indexes are still the clear winner, but only between 18 and 25 percent faster.

For collections, iterators are faster than indexes

For a List of Integers, however, iterators are the clear winner. Just change the int-array in the test-class to:

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});

And make the necessary changes to the test-function (int[] to List<Integer>, length to size(), etc.):

[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)

In one test they're almost equivalent, but with collections, iterator wins.

*This post is based on two answers I wrote on Stack Overflow:

Uses and syntax for for-each loop in Java
Should I use an Iterator or a forloop to iterate?

Some more information: Which is more efficient, a for-each loop, or an iterator?

The full testing class

I created this compare-the-time-it-takes-to-do-any-two-things class after reading this question on Stack Overflow:

import  java.text.NumberFormat;
import  java.util.Locale;

/**
   &lt;P&gt;{@code java TimeIteratorVsIndexIntArray 1000000}&lt;/P&gt;

   @see  &lt;CODE&gt;&lt;A HREF=&quot;https://stackoverflow.com/questions/180158/how-do-i-time-a-methods-execution-in-java&quot;&gt;https://stackoverflow.com/questions/180158/how-do-i-time-a-methods-execution-in-java&lt;/A&gt;&lt;/CODE&gt;
 **/
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 a 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

}

Here is an answer which does not assume knowledge of Java Iterators. It is less precise, but it is useful for education.

While programming we often write code that looks like the following:

char[] grades = ....
for(int i = 0; i < grades.length; i++) {   // for i goes from 0 to grades.length
    System.out.print(grades[i]);           // Print grades[i]
}

The foreach syntax allows this common pattern to be written in a more natural and less syntactically noisy way.

for(char grade : grades) {   // foreach grade in grades
    System.out.print(grade); // print that grade
}

Additionally this syntax is valid for objects such as Lists or Sets which do not support array indexing, but which do implement the Java Iterable interface.

The for-each loop in Java uses the underlying iterator mechanism. So it's identical to the following:

Iterator<String> iterator = someList.iterator();

while (iterator.hasNext()) {
  String item = iterator.next();
  System.out.println(item);
}