masking zero columns in numpy.array

Question

I'm trying to do the following, just misunderstand where I'm wrong:

In: import numpy as np

In: a  = np.array([[0., 1., 0.],
             [1., 2., 0.],
             [0., 3., 0.]]) 

In: for i in range(a.shape[1]):
        a[:, i] = np.ma.masked_where(~a[:, i].any(), a[:, i])

Out: array([[0., 1., --],
     [1., 2., --],
     [0., 3., --]])

The point here is to mask only the third column because it's completely zeros, and leave the first column zeros unmasked

@Alex -- is what NumPy will output when a value is masked. — user707650
– user707650, Commented Jan 22, 2018 at 22:00
You haven't asked an actual question, but I assume your question is: why are there (still) two non-masked zeros in the final output. But please update your question with your actual question. — user707650
– user707650, Commented Jan 22, 2018 at 22:02

user707650 · Accepted Answer · 2018-01-22 22:07:10Z

1

You can't invert (~) floating point numbers and arrays.

Try with:

a = np.ma.masked_where(np.isclose(a, 0), a)

(no for loop needed either.)

answered Jan 22, 2018 at 22:07

user707650

Sign up to request clarification or add additional context in comments.

Comments

HFBrowning · Accepted Answer · 2018-01-23 19:28:26Z

0

One approach to mask based on all zeros in a column is simply to identify column sums of zero. Below is my solution, borrowing very heavily off of this SO answer to provide a simple mechanism for masking after you have gotten the column indices of interest.

import numpy as np    

def mask_zero_cols(in_array):
    # idx is an array of column indices for all-zero columns
    idx = np.where(in_array.sum(axis=0) == 0)[0] #axis=0 is for columns
    m = np.zeros_like(in_array)
    m[:, idx] = 1

    return np.ma.masked_array(in_array, m)


a = np.array([[0., 1., 0.],
             [1., 2., 0.],
             [0., 3., 0.]])

a_mask = mask_zero_cols(a)

edited Jan 23, 2018 at 19:28

answered Jan 22, 2018 at 21:57

HFBrowning

2,3844 gold badges25 silver badges45 bronze badges

Collectives™ on Stack Overflow

masking zero columns in numpy.array

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related