do we really want to support using multiple columns? Imho, yes. Do we support complete rows? How?
what convenience functions should be offered on the front-end side? E.g. getters allowing for using of col names and indices.
Are the indices in getters the indices in the referenced DataFrame, or is front-end indexing relative to the number of indices used for the columnDimension?
things that need to be done:
implement tests
update docstrings (some have not been updated to reflect the option of multiple columns)
documentation and examples for read the docs
Wiki update regarding columnDimension and DataFrame
Not entirely done yet, e.g. tests are missing. But please feel free to comment.