Strange behaviour when subsetting FlowDataFrame

[mauruscz]

If I have a FlowDataFrame whose columns are

origin | destination | flow | attr_1 | attr_2

and I try to take a subset of column, it is only allowed when I only pick the origin OR the destination

Like

fdf[["origin", "flow"]] or fdf[["destination", "flow"]]

But if I do something like fdf["origin", "destination", "flow", "attr_1"] it fails with the error

File [/opt/anaconda3/envs/GSA/lib/python3.9/site-packages/skmob/core/flowdataframe.py:319](https://file+.vscode-resource.vscode-cdn.net/opt/anaconda3/envs/GSA/lib/python3.9/site-packages/skmob/core/flowdataframe.py:319), in FlowDataFrame.__getitem__(self, key)
    314 """
    315 It the result contains lat, lng and datetime, return a TrajDataFrame, else a pandas DataFrame.
    316 """
    317 result = super(FlowDataFrame, self).__getitem__(key)
--> 319 if (isinstance(result, FlowDataFrame)) and result._is_flowdataframe():
    320     result.__class__ = FlowDataFrame
    321     result.tessellation = self._tessellation

File [/opt/anaconda3/envs/GSA/lib/python3.9/site-packages/skmob/core/flowdataframe.py:277](https://file+.vscode-resource.vscode-cdn.net/opt/anaconda3/envs/GSA/lib/python3.9/site-packages/skmob/core/flowdataframe.py:277), in FlowDataFrame._is_flowdataframe(self)
    271 def _is_flowdataframe(self):
    273     if ((constants.ORIGIN in self) and
    274             pd.core.dtypes.common.is_string_dtype(self[constants.ORIGIN])) \
    275         and ((constants.DESTINATION in self) and
    276              pd.core.dtypes.common.is_string_dtype(self[constants.DESTINATION])) \
--> 277         and ((constants.TILE_ID in self._tessellation) and
    278              pd.core.dtypes.common.is_string_dtype(self._tessellation[constants.TILE_ID])) \
    279         and ((constants.FLOW in self) and
    280              (pd.core.dtypes.common.is_float_dtype(self[constants.FLOW]) or
    281               pd.core.dtypes.common.is_integer_dtype(self[constants.FLOW]))):
    282         return True
    284     return False

TypeError: argument of type 'NoneType' is not iterable