[python] How to delete columns in numpy.array

I would like to delete selected columns in a numpy.array . This is what I do:

n [397]: a = array([[ NaN,   2.,   3., NaN],
   .....:        [  1.,   2.,   3., 9]])

In [398]: print a
[[ NaN   2.   3.  NaN]
 [  1.   2.   3.   9.]]

In [399]: z = any(isnan(a), axis=0)

In [400]: print z
[ True False False  True]

In [401]: delete(a, z, axis = 1)
Out[401]:
 array([[  3.,  NaN],
       [  3.,   9.]])

In this example my goal is to delete all the columns that contain NaN's. I expect the last command to result in:

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

How can I do that?

Example from the numpy documentation:

>>> a = numpy.array([[ 0,  1,  2,  3],
               [ 4,  5,  6,  7],
               [ 8,  9, 10, 11],
               [12, 13, 14, 15]])

>>> numpy.delete(a, numpy.s_[1:3], axis=0)                       # remove rows 1 and 2

array([[ 0,  1,  2,  3],
       [12, 13, 14, 15]])

>>> numpy.delete(a, numpy.s_[1:3], axis=1)                       # remove columns 1 and 2

array([[ 0,  3],
       [ 4,  7],
       [ 8, 11],
       [12, 15]])

>>> A = array([[ 1,  2,  3,  4],
               [ 5,  6,  7,  8],
               [ 9, 10, 11, 12]])

>>> A = A.transpose()

>>> A = A[1:].transpose()

In your situation, you can extract the desired data with:

a[:, -z]

"-z" is the logical negation of the boolean array "z". This is the same as:

a[:, logical_not(z)]

This creates another array without those columns:

  b = a.compress(logical_not(z), axis=1)

From Numpy Documentation

np.delete(arr, obj, axis=None) Return a new array with sub-arrays along an axis deleted.

>>> arr
array([[ 1,  2,  3,  4],
       [ 5,  6,  7,  8],
       [ 9, 10, 11, 12]])
>>> np.delete(arr, 1, 0)
array([[ 1,  2,  3,  4],
       [ 9, 10, 11, 12]])

>>> np.delete(arr, np.s_[::2], 1)
array([[ 2,  4],
       [ 6,  8],
       [10, 12]])
>>> np.delete(arr, [1,3,5], None)
array([ 1,  3,  5,  7,  8,  9, 10, 11, 12])

Removing Matrix columns that contain NaN. This is a lengthy answer, but hopefully easy to follow.

def column_to_vector(matrix, i):
    return [row[i] for row in matrix]
import numpy
def remove_NaN_columns(matrix):
    import scipy
    import math
    from numpy import column_stack, vstack

    columns = A.shape[1]
    #print("columns", columns)
    result = []
    skip_column = True
    for column in range(0, columns):
        vector = column_to_vector(A, column)
        skip_column = False
        for value in vector:
            # print(column, vector, value, math.isnan(value) )
            if math.isnan(value):
                skip_column = True
        if skip_column == False:
            result.append(vector)
    return column_stack(result)

### test it
A = vstack(([ float('NaN'), 2., 3., float('NaN')], [ 1., 2., 3., 9]))
print("A shape", A.shape, "\n", A)
B = remove_NaN_columns(A)
print("B shape", B.shape, "\n", B)

A shape (2, 4) 
 [[ nan   2.   3.  nan]
 [  1.   2.   3.   9.]]
B shape (2, 2) 
 [[ 2.  3.]
 [ 2.  3.]]

Another way is to use masked arrays:

import numpy as np
a = np.array([[ np.nan,   2.,   3., np.nan], [  1.,   2.,   3., 9]])
print(a)
# [[ NaN   2.   3.  NaN]
#  [  1.   2.   3.   9.]]

The np.ma.masked_invalid method returns a masked array with nans and infs masked out:

print(np.ma.masked_invalid(a))
[[-- 2.0 3.0 --]
 [1.0 2.0 3.0 9.0]]

The np.ma.compress_cols method returns a 2-D array with any column containing a masked value suppressed:

a=np.ma.compress_cols(np.ma.masked_invalid(a))
print(a)
# [[ 2.  3.]
#  [ 2.  3.]]

See manipulating-a-maskedarray