Take only part of matrix in Python

Question

given a matrix A, e.g. A = np.array([[1,2,3,4],[5,6,7,8]]) and two lists one of row indices and one of column indices, e.g. row=(0,1), column=(0,2) I want to extract the corresponding rows and columns of matrix A in python, so in this example I want the result to be np.array([[1, 3], [5,7])). I know how to adress a single entry in a matrix, but not all within a list, moreover I am always loosing the structure. So far the best result I got was with A[row, column], which does not return all indices listed, only A=np.array(([1,7])). I also know that there exists a slicing command, but this only works for consecutive rows and columns which is not the case.

Thank you very much in advance!

score 2 · Accepted Answer · answered Jun 13 '22 at 11:01

2

It looks like you want:

A[np.array(row)[:,None], column]

output:

array([[1, 3],
       [5, 7]])

intermediate:

np.array(row)[:,None]

array([[0],
       [1]])

answered Jun 13 '22 at 11:01

mozway

194,879
13
39
75

Also, it's possible to use `np.newaxis`: `A[row[:, np.newaxis], column]` – Gil Ben David Jun 13 '22 at 11:05
1

np.newaxis is just an alias for None (to make axis expansion more readable) – Learning is a mess Jun 13 '22 at 12:00
Yes, that's exactly what I wanted. Thanks! – samabu Jun 13 '22 at 12:06

score 1 · Answer 2 · answered Jun 13 '22 at 11:00

1

np.array([[A[r][c] for c in column] for r in row]) # array([[1, 3], [5, 7]])

answered Jun 13 '22 at 11:00

remarcoble

537
1
4
19

1

two drawbacks, as you loop this will be slow on large arrays, and you cannot use this to index/assign the existing array, this creates a new one ;) – mozway Jun 13 '22 at 11:04

Take only part of matrix in Python

2 Answers2