UnivariateFitter subtract/divide return all NaNs when a label is specified

[yanirs]

For example, running this code:

from lifelines import KaplanMeierFitter
from lifelines.datasets import load_waltons

df = load_waltons()
ix = df['group'] == 'miR-137'

kmf1 = KaplanMeierFitter()
kmf1.fit(df['T'][~ix], df['E'][~ix], label='control')

kmf2 = KaplanMeierFitter()
kmf2.fit(df['T'][ix], df['E'][ix], label='miR-137')

kmf1.subtract(kmf2).head()

yields this DataFrame:

timeline	control	miR-137
0.0	NaN	NaN
6.0	NaN	NaN
7.0	NaN	NaN
9.0	NaN	NaN
13.0	NaN	NaN

But if we drop the labels, we get:

timeline	KM_estimate
0.0	0.000000
6.0	0.029412
7.0	0.021660
9.0	0.109895
13.0	0.198130

Shouldn't the label be ignored by subtract/divide?

[CamDavidsonPilon]

Ah, that's a legit bug. Thanks for reporting!

[CamDavidsonPilon]

This was fixed in the latest version

[yanirs]

Woohoo! Thanks!