Efficiently create dummies of all factors and character vectors in a data frame. Support is included for learning the categories on one data set (e.g., a training set) and deploying them on another (e.g., a test set).
