Series/DataFrame.rank() doesn't handle np.inf properly

[nspies]

The rank() function converts np.nan to np.inf or -np.inf, and therefore can't properly rank existing np.inf. See here: #6886 (comment).

Test for this issue here: https://github.com/pydata/pandas/blob/master/pandas/tests/test_series.py#L4071

[jreback]

@nspies this has your name on it too!

[peterpanmj]

I'm using pandas '0.20.1'. Ranking series with np.inf still yields np.nan. Will this be fixed ?

s= pd.Series(data = np.array([1,2,3,math.inf,4,4,4,5,5]))
s.rank(method="average")

0    1.0
1    2.0
2    3.0
3    NaN
4    5.0
5    5.0
6    5.0
7    7.5
8    7.5
dtype: float64

[jreback]

@peterpanmj pull request for this are welcome, its pretty straightforward to fix.