filtering entries of HashMap and iterate over values

Question

I have useed a HashMap to implement a graph

HashMap<E,Node<E>> graph=new HashMap<E,Node<E>>()

now I want to filter the head nodes out of it (Node class has a method that shows if this node is head or tail)

public Iterator<Node<E>> heads() {
    HashMap<E, Node<E>> heads=new HashMap<E, Node<E>>();
    for(Entry<E, Node<E>> e:graph.entrySet()){
        if(e.getValue().isHead())
            heads.put(e.getKey(), e.getValue());
    }
    return heads.values().iterator();
}

I wanted to know if it is a good way of implementation or is there a better proper way ?

Answer 1

Don't know in what aspect of "Better" you are looking for. Your way works anyway.

There is another way I prefer, by using Guava's Iterables .

it looks something like:

Iterable<Node<E>> heads 
     = Iterables.filter(graph.values(), 
                        new Predicate<Node<E>> {
                            @Override
                            public boolean apply(Node<E> node) {
                                return node.isHead();
                            }
                       );

---

Edit: If you are using Java 8, you will already find similar feature without 3rd party lib. Thanks to Lambda, the syntax is even easier. Anyway, the idea is the same.

Answer 2

One straightforward improvement over the algorithm you have posted would be to collect the head elements of the map into a list, rather than reconstructing a HashMap with just the head elements in it:

public Iterator<Node<E>> heads() {
    List<Node<E>> heads = new ArrayList<>();
    for (Node<E> e : graph.values()) {
        if (e.isHead())
            heads.add(e);
    }
    return heads.iterator();
}

Answer 3

Java 8版本：

Iterator<Node> s = graph.values().stream().filter(n -> n.isHead()).iterator();