BUG: astype('Int64') raises AttributeError

[jorisvandenbossche]

In [8]: df = pd.DataFrame([[1., 2.], [3., 4.], [5., 6.]], columns=['a', 'b'])

In [9]: df
Out[9]: 
     a    b
0  1.0  2.0
1  3.0  4.0
2  5.0  6.0

In [10]: df.astype('Int64')
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-10-b9d2763e69d8> in <module>()
----> 1 df.astype('Int64')

~/scipy/pandas/pandas/util/_decorators.py in wrapper(*args, **kwargs)
    175                 else:
    176                     kwargs[new_arg_name] = new_arg_value
--> 177             return func(*args, **kwargs)
    178         return wrapper
    179     return _deprecate_kwarg

~/scipy/pandas/pandas/core/generic.py in astype(self, dtype, copy, errors, **kwargs)
   5162             # else, only a single dtype is given
   5163             new_data = self._data.astype(dtype=dtype, copy=copy, errors=errors,
-> 5164                                          **kwargs)
   5165             return self._constructor(new_data).__finalize__(self)
   5166 

~/scipy/pandas/pandas/core/internals/managers.py in astype(self, dtype, **kwargs)
    554 
    555     def astype(self, dtype, **kwargs):
--> 556         return self.apply('astype', dtype=dtype, **kwargs)
    557 
    558     def convert(self, **kwargs):

~/scipy/pandas/pandas/core/internals/managers.py in apply(self, f, axes, filter, do_integrity_check, consolidate, **kwargs)
    421 
    422             kwargs['mgr'] = self
--> 423             applied = getattr(b, f)(**kwargs)
    424             result_blocks = _extend_blocks(applied, result_blocks)
    425 

~/scipy/pandas/pandas/core/internals/blocks.py in astype(self, dtype, copy, errors, values, **kwargs)
    562     def astype(self, dtype, copy=False, errors='raise', values=None, **kwargs):
    563         return self._astype(dtype, copy=copy, errors=errors, values=values,
--> 564                             **kwargs)
    565 
    566     def _astype(self, dtype, copy=False, errors='raise', values=None,

~/scipy/pandas/pandas/core/internals/blocks.py in _astype(self, dtype, copy, errors, values, klass, mgr, **kwargs)
    679                     "current ({newb_dtype} [{newb_size}])".format(
    680                         copy=copy, dtype=self.dtype.name,
--> 681                         itemsize=self.itemsize, newb_dtype=newb.dtype.name,
    682                         newb_size=newb.itemsize))
    683         return newb

AttributeError: 'FloatBlock' object has no attribute 'itemsize'

The same if the starting dataframe is of integer dtype.