Hi I would like to consider the option of converting the underlying structure of a sparse Dataframe to a sparse matrix instead of multiple sparse series in case all the columns of the Dataframe are of the same dtype. This will make row/column slicing of the Dataframe much faster that what it is currently.