Search for a specific row and remove from the array in Python-CodePudding

I have an array A. I want to seek rows with all zeroes and then remove them. Here I want to remove A[:,2] and append the row index to B. The desired output is attached.

import numpy as np

A=np.array([[ 0.00000000e 00,  3.57765318e-08,  0.00000000e 00,
         1.74215085e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00],
       [-1.06733099e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  3.50573100e-08,  0.00000000e 00,
         0.00000000e 00],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00],
       [-1.78530448e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  1.61650808e-08,  9.06895783e-08,
         0.00000000e 00],
       [ 0.00000000e 00, -4.46583743e-08,  0.00000000e 00,
        -5.99482549e-08,  0.00000000e 00,  0.00000000e 00,
         1.04606629e-07],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
        -9.64183682e-08,  0.00000000e 00,  0.00000000e 00,
         9.64183682e-08],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00, -1.25633650e-07, -4.03926936e-07,
         0.00000000e 00]])

B=[5,8]

The desired output is

array([[ 0.00000000e 00,  3.57765318e-08,  0.00000000e 00,
         1.74215085e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00],
       [-1.06733099e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  3.50573100e-08,  0.00000000e 00,
         0.00000000e 00],
       [-1.78530448e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  1.61650808e-08,  9.06895783e-08,
         0.00000000e 00],
       [ 0.00000000e 00, -4.46583743e-08,  0.00000000e 00,
        -5.99482549e-08,  0.00000000e 00,  0.00000000e 00,
         1.04606629e-07],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
        -9.64183682e-08,  0.00000000e 00,  0.00000000e 00,
         9.64183682e-08],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00, -1.25633650e-07, -4.03926936e-07,
         0.00000000e 00]])
B=[2,5,8]

CodePudding user response：

You can use:

# find all rows where all values are 0
x = np.where((A==0).all(1))[0]

# delete them from A
A = np.delete(A, x, axis=0)

# prepend them in B
B = np.r_[x, B]

Alternative:

mask = (A==0).all(1)
x = np.where(mask)[0]

A = A[~mask]

B = np.r_[x, B]

output:

# A
array([[ 0.00000000e 00,  3.57765318e-08,  0.00000000e 00,
         1.74215085e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00],
       [-1.06733099e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  3.50573100e-08,  0.00000000e 00,
         0.00000000e 00],
       [-1.78530448e-07,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00,  1.61650808e-08,  9.06895783e-08,
         0.00000000e 00],
       [ 0.00000000e 00, -4.46583743e-08,  0.00000000e 00,
        -5.99482549e-08,  0.00000000e 00,  0.00000000e 00,
         1.04606629e-07],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
        -9.64183682e-08,  0.00000000e 00,  0.00000000e 00,
         9.64183682e-08],
       [ 0.00000000e 00,  0.00000000e 00,  0.00000000e 00,
         0.00000000e 00, -1.25633650e-07, -4.03926936e-07,
         0.00000000e 00]])

# B
array([2, 5, 8])

CodePudding user response：

IIUC, you can do something like:

# the index of the first matching row
x = np.all(A==0, axis=1).argmax()

A_bar = np.delete(A, x, axis=0)
B.append(x)

CodePudding user response：

Here's a way:

C = np.all(A==0, axis=1)
D = np.arange(A.shape[0])[C]
A = A[~C, :]
print(A)
B.extend(D)
print(B)

Output:

[[ 0.00000000e 00  3.57765318e-08  0.00000000e 00  1.74215085e-07
   0.00000000e 00  0.00000000e 00  0.00000000e 00]
 [-1.06733099e-07  0.00000000e 00  0.00000000e 00  0.00000000e 00
   3.50573100e-08  0.00000000e 00  0.00000000e 00]
 [-1.78530448e-07  0.00000000e 00  0.00000000e 00  0.00000000e 00
   1.61650808e-08  9.06895783e-08  0.00000000e 00]
 [ 0.00000000e 00 -4.46583743e-08  0.00000000e 00 -5.99482549e-08
   0.00000000e 00  0.00000000e 00  1.04606629e-07]
 [ 0.00000000e 00  0.00000000e 00  0.00000000e 00 -9.64183682e-08
   0.00000000e 00  0.00000000e 00  9.64183682e-08]
 [ 0.00000000e 00  0.00000000e 00  0.00000000e 00  0.00000000e 00
  -1.25633650e-07 -4.03926936e-07  0.00000000e 00]]
[5, 8, 2]

CodePudding user response：

You can do this as:

mask = (A == 0).all(1)
B = np.append(B, np.arange(A.shape[0])[mask])  # .tolist()
A = A[~mask]