quantile (and describe) not ignoring missing values with pandas 0.18.1?

[Etienne Chamayou]

Hi,

I've just run some code, written while using pandas 0.15.2, with pandas 0.18.1, and with the latter missing values are not ignored by quantile so that:

pd.Series([1,2,3,np.nan]).quantile(0.5) returns nan 

pd.Series([1,2,3,np.nan]).describe() returns nan for each percentile

Is it normal? (If yes, why not, the doc of quantile is ok (use of numpy.percentile) but for describe (generate various summary statistics, excluding NaN value) I'm not so sure.)

Cheers,

Etienne

[Joris Van den Bossche]

Hi Etienne,

This was a regression in 0.18.1, and which will be fixed in the next release (see https://github.com/pydata/pandas/issues/13098)

Regards,

Joris

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.