如何使用Iterator从集合中删除元素？

Question

I have a scenario that I am iterating over a set using iterator. Now I want to remove 1st element while my iterator is on 2nd element. How can I do it. I know Set is unordered and there is nothing like first or second element but my question is I want to remove an element which is not being currently returned by Iterator.next

1. I dont want to convert this set to list and using listIterator.

2. I dont want to collect all objects to be removed in other set and call remove all

3. I cant store it and remove after the iteration

sample code.

Set<MyObject> mySet = new HashSet<MyObject>();

mySet.add(MyObject1);
mySet.add(MyObject2);

...
Iterator itr = mySet.iterator();
        while(itr.hasNext())
        {
            // Now iterator is at second element and I want to remove first element
        }

Answer 1

HashSet is unordered and javadoc clearly states that Iterator's remove method Removes from the underlying collection the last element returned by this iterator (optional operation). So the answer is no through an iterator.Since HashSet contains unique elements,you can use Set.remove(Object) after traversing the first element,in thios case you dont even need to go to the 2nd element

HashSet<K> hs;// you HashSet containing unique elements

if(!hs.isEmpty())
{
  hs.remove(hs.iterator().next());
}

Just remember HashSet is unordered and there is no such thing as 1st or 2nd element

Alternately,you should use LinkedHashSet which gives you an ordered Set based on insertion order

Answer 2

Set.iterator() returns a java.lang.Iterator . This iterator only provides methods to remove the current element and to iterate forward .

So if you don't want to convert your set, using only Iterator you cannot remove the previous element.

What you can do for example is that you collect the elements you want to remove, and after you iterated through the whole set, you remove the collected elements after, eg with Set.removeAll(removableCollection) :

List<MyObject> removableList = new ArrayList<>();

MyObject previous;

Iterator<MyObject> itr = mySet.iterator();
while (itr.hasNext()) {
    MyObject current = itr.next();

    // If you find you want to remove the previous element:
    if (someCondition)
        removableList.add(previous);

    previous = current;
}

mySet.removeAll(removeableList);

Answer 3

Given the constraints as you have stated them, I don't think that there is solution to the problem.

• The Iterator.remove() method will only remove the "current" element.
• You have excluded "remembering" objects and removing them from the HashSet in a second pass / phase.
• Schemes that involve using two iterators simultaneously and removing using one of them will result in CCMEs on the second one.

The three approaches that you suggested (but then excluded) would all work. I think the 2nd one would be the most performant.

Another idea would be to implement a new hash table-based Set type which has an Iterator with an extra remove operation. (You could start with the source code of HashSet etcetera, rename the class and then modify it to do what you need.)