LibGDX Using a Map with Duplicate keys vs using instanceof

Game Mechanics
#1

I’m creating an Array to hold a number of objects which all implement a given interface. Much of the time, it won’t matter what the subtype of the Object is, but, occasionally it will and I want to add some methods to get those subtypes for when required eg getA(), getB() etc.

I’m thinking that indexing them by a number for each subtype would be the quickest way to do this, although it would require a collection class that allows for duplicate keys:


private Array<SuperInterface> objects;

public Array<A> getA() {
    return objects.get(0); //0 being the index number for objects of type A
}

the alternative is simply to iterate through and create a new Array of that type:


private Array<SuperInterface> objects;

public Array<A> getA() {
    Array<A> a = new Array<A>();
    for (SuperInterface super : objects) {
        if (super instanceof A) {
            a.add(super);
        }
    }
    return a;

This, however, seems to me to be the least efficient approach.

So, my questions are:

  1. Do you agree with my assessment regarding efficiency? If not, what would you suggest?

  2. Are you aware of a good Map implementation that allows duplicate keys?

I’m aware that Google’s Guava includes a MultiMap implementation, but it seems overkill to include a library simply to use one class.

Hope that all makes sense!