pandas.Series.get fails with: object has no attribute 'values'

Question

I don't seem to be able to call Series.get on a Series object.

>> print col
0    1
1    1
2    0
Name: a, dtype: float64
>>> counts = col.value_counts()
>>> print counts
1    2
0    1
dtype: int64

... makes sense. 2 ones. 1 zero

>>> print type(counts)
<class 'pandas.core.series.Series'>

... OK. The result is a Series. How can I read out the elements? According to Series.get , and the docstring for counts.get, I should be able to:

zeros = counts.get(0,0)
ones = counts.get(1,0)

... but this fails with:

AttributeError: 'numpy.ndarray' object has no attribute 'values'

What I have misunderstood?

>>> help(counts.get)
Help on method get in module pandas.core.series:

get(self, label, default=None) method of pandas.core.series.Series instance
    Returns value occupying requested label, default to specified
    missing value if not present. Analogous to dict.get

    Parameters
    ----------
    label : object
        Label value looking for
    default : object, optional
        Value to return if label not in index

    Returns
    -------
    y : scalar

In:

>>> print counts
1    2
0    1

aren't 1 and 0 the labels?

Answer 1

Seems to be bug in pandas 15

Upgrading to pandas 18 resolves this.