Alright, here we go. After a long long long time I have decided to yet again try and maintain a blog.

The other day at work I was trying to use .NET’s Dictionary class to create a hierarchical structure which allows me to do something like this:

torqueInfo[eAvatarGender.male][eJointName.elbow][eJointDirection.extension]
	= new TorqueCurves(eAvatarGender.male, eJointName.elbow, eJointDirection.extension);

torqueInfo has been defined as:

Dictionary<eAvatarGender, Dictionary<eJointName, Dictionary<eJointDirection, TorqueCurves>>> torqueInfo;

Using the dictionary object this way would help me reduce my coding clutter to just 3 look-ups instead of coding a 3 level conditional switch (one of the straightforward ways of doing it).

The problem with this is that unlike C++ std::map objects, if a .NET dictionary doesn’t contain the specified key it throws an exception. 

Dictionary<int, string> keyVal = new Dictionary<int, string> ();
keyVal [5] = "Hello";		// exception
keyVal.Add (5, "Hello");	// OK

Filling up such a construct with an Add method could have easily turned into a nightmare. To tackle this problem, I created my own little dictionary class which could handle unknown keys and allow the programmer to easily populate a complex construct like I mentioned above. Here’s what the class looks like:

public class SmarterDictionary<TKey, TValue> where TValue : new ()
{
    Dictionary<TKey, TValue> keyVal = new Dictionary<TKey,TValue> ();

    // access operator is all we need!
    public TValue this[TKey key] {
        get {
            if (!keyVal.ContainsKey(key))
                keyVal.Add(key, new TValue ());

            return keyVal[key];
        }
        set {
            keyVal[key] = value;
        }
    }
}

HTH