data process dictionary for quick look up needed?

[DaweiPang32]

          is this necessary? With pandas, if you set one (or multiple) column(s) as the index, then it's already kinda a hashmap (i.e. a python dictionary-ish)? It seems to me that doing df.iterrows() would bring a huge performance hit.

This might be a bigger change. feel free to create a create an issue to track this and leave this for later.

Originally posted by @chunhochow in #18 (comment)