[FIXED] How can I drop columns with zero variance from a numpy array?

Issue

So using the following example:

import numpy as np
X = np.array([[1,10,np.nan,0],[2,10,np.nan,0],[3,10,np.nan,0]])

I can drop the all-NaN or all-0 columns like so:

 X = X[:,~np.all(np.isnan(X), axis=0)]
 X = X[:,~np.all(X==0, axis=0)]

I would also like to drop the zero variance columns, is there a good way to do so using the same concise notation?

Thanks!

Solution

You can.

import numpy as np

X[:, np.var(X, axis=0) != 0]

array([[ 1., nan],
       [ 2., nan],
       [ 3., nan]])

This calculates the variance of each column (axis=0) and then returns the ones not equal to 0.

Answered By - m13op22