The iterator semantics for Trove's primitive maps is slightly different from those defined in java.util.Iterator, but still well within the scope of the pattern, as defined by Gamma, et al.

This iterator does not implicitly advance to the next entry when the value at the current position is retrieved. Rather, you must explicitly ask the iterator to advance() and then retrieve either the key(), the value() or both. This is done so that you have the option, but not the obligation, to retrieve keys and/or values as your application requires, and without introducing wrapper objects that would carry both. As the iteration is stateful, access to the key/value parts of the current map entry happens in constant time.

In practice, the iterator is akin to a "search finger" that you move from position to position. Read or write operations affect the current entry only and do not assume responsibility for moving the finger.

Here are some sample scenarios for this class of iterator:

 // accessing keys/values through an iterator:
 for ( TObjectLongIterator it = map.iterator(); it.hasNext(); ) {
   it.advance();
   if ( satisfiesCondition( it.key() ) ) {
     doSomethingWithValue( it.value() );
   }
 }
 

 // modifying values in-place through iteration:
 for ( TObjectLongIterator it = map.iterator(); it.hasNext(); ) {
   it.advance();
   if ( satisfiesCondition( it.key() ) ) {
     it.setValue( newValueForKey( it.key() ) );
   }
 }
 

 // deleting entries during iteration:
 for ( TObjectLongIterator it = map.iterator(); it.hasNext(); ) {
   it.advance();
   if ( satisfiesCondition( it.key() ) ) {
     it.remove();
   }
 }
 

 // faster iteration by avoiding hasNext():
 TObjectLongIterator iterator = map.iterator();
 for ( int i = map.size(); i-- > 0; ) {
   iterator.advance();
   doSomethingWithKeyAndValue( iterator.key(), iterator.value() );
 }