Support pandas objects directly in the Dataset constructor?

[shoyer]

Currently, this fails:

In [32]: xray.Dataset(df)
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-32-358846bebb6d> in <module>()
----> 1 xray.Dataset(df)

/Users/shoyer/dev/xray/xray/core/dataset.pyc in __init__(self, data_vars, coords, attrs, compat, **kwargs)
    206         if coords is None:
    207             coords = set()
--> 208         if data_vars or coords:
    209             self._set_init_vars_and_dims(data_vars, coords, compat)
    210         if attrs is not None:

/Users/shoyer/miniconda/envs/chinook-py2/lib/python2.7/site-packages/pandas/core/generic.pyc in __nonzero__(self)
    712         raise ValueError("The truth value of a {0} is ambiguous. "
    713                          "Use a.empty, a.bool(), a.item(), a.any() or a.all()."
--> 714                          .format(self.__class__.__name__))
    715
    716     __bool__ = __nonzero__

ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

But there's no reason why this shouldn't work. After #671, this is as simple as converting the DataFrame or Series to an OrderedDict first:

In [31]: xray.Dataset(OrderedDict(df))
Out[31]:
<xray.Dataset>
Dimensions:  (x: 3)
Coordinates:
  * x        (x) object 'a' 'b' 'c'
Data variables:
    y        (x) int64 0 1 2

In [34]: xray.Dataset(OrderedDict(df.y))
Out[34]:
<xray.Dataset>
Dimensions:  ()
Coordinates:
    *empty*
Data variables:
    a        int64 0
    b        int64 1
    c        int64 2

[max-sixty]

@shoyer I think this can be closed now?

[shoyer]

Yes. Just added #740 so we don't forget about Series.