Series.rolling(3).quantile(10) segmentation fault

[kernc]

Code Sample, a copy-pastable example if possible

>>> pd.Series(np.arange(10)).rolling(3).quantile(10)
[1]    20214 segmentation fault

Problem description

The value 10 is over the fraction of 1 which .quantile() works with, but I guess it still shouldn't crash my interpreter.

Expected Output

ValueError: quantile value 10 not in [0, 1]

Output of pd.show_versions()

pandas 0.19.0+479.g4842bc7

[jreback]

yep, not very friendly...PR's appreciated!

[csizsek]

Hi @kernc and @jreback , I'm thinking about fixing this issue. Any idea where to start? (I have read the contribution documentation, set up the environment and was able to reproduce.)

[jreback]

you can add a validation right here 0 <= quantile <= 1: https://github.com/pandas-dev/pandas/blob/master/pandas/window.pyx#L1288, and raise a ValueError if not in range. please add some tests as well.

thanks!