Efficiently test matrix rows and columns with numpy

Question

I am trying to remove both the row i and column i when both the row i and column i contains all 0s. For example in this case we can see that row 0 is all zeros

Answer 1

Vectorized approach with masking -

def remove_vectorized(a):
    mask = a==0
    m_row = ~mask.all(1)
    m_col = ~mask.all(0)
    comb_mask = m_row | m_col
    return a[comb_mask][:,comb_mask] #or a[np.ix_(comb_mask, comb_mask)]

Sample runs

Case #1 :

In [485]: a
Out[485]: 
array([[0, 0, 0, 0, 0],
       [0, 1, 0, 1, 0],
       [0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0]])

In [486]: remove_vectorized(a)
Out[486]: 
array([[1, 1],
       [0, 0]])

Case #2 :

In [489]: a
Out[489]: 
array([[0, 0, 1, 0, 0, 1],
       [0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0],
       [0, 0, 1, 0, 1, 0]])

In [490]: remove_vectorized(a)
Out[490]: 
array([[0, 1, 0, 1],
       [0, 0, 0, 0],
       [0, 0, 0, 0],
       [0, 1, 1, 0]])