Data.xnew() is broken

[chrkorn]

The assignment to df.values does not work and leaves all feature data set to 0.

[clementpoiret]

I encountered the same problem with the following code:

import numerox as nx
from sklearn.preprocessing import StandardScaler

data = nx.load_zip('.numerai_datasets.zip', single_precision=True)
sc = StandardScaler()
x = sc.fit_transform(data.x)
data = data.xnew(x)

data.x is now a matrix full of zeros

EDIT: when setting single_precision=False, xnew has the correct behavior