storing a dataframe in hdf5 with table= True and a hierarchical index

[Arthur Gerigk]

this is currently not possible. (message below).

is this a bug or is there some technical restiction that prevents this from being possible?

/usr/local/lib/python2.7/dist-packages/tables/description.pyc in __init__(self, classdict, nestedlvl, validate)
    464   subclass) instance and got: "%s". Please make use of the Col(), or
    465   descendant, constructor to properly initialize columns.
--> 466 """ % object
    467             object._v_pos = pos  # Set the position of this object
    468             object._v_parent = self  # The parent description

TypeError: Passing an incorrect value to a table column. Expected a Col (or
  subclass) instance and got: "ObjectAtom()". Please make use of the Col(), or
  descendant, constructor to properly initialize columns.

[Wes McKinney]

I think the answer is that it's "neither" a bug nor a technical
limitation; simply no support for hierarchical indexes has been built
into the table format yet. Creating an issue here to keep track of it:

https://github.com/pydata/pandas/issues/1277