convert_objects incorrectly converting bools to nans

[cpcloud]

SO xref: http://stackoverflow.com/q/23658092/564538

In [45]: paste
>>> s1 = Series([1, True, 3, 5], index=['a', 'b', 'c', 'e'])

## -- End pasted text --

In [46]: s1
Out[46]:
a       1
b    True
c       3
e       5
dtype: object

In [47]: s1.convert_objects(convert_numeric=True)
Out[47]:
a     1
b   NaN
c     3
e     5
dtype: float64

[jreback]

you mean you think the bools should be coerced to integers? (e.g. 1 in this case), as opposed to eliminated any non-integer/float values; makes sense

[cpcloud]

yep they should be cast to the type of the surrounding elements