Select using multiple arrays of indices from an ndarray in numpy

Question

Say if I have a 2D array:

y = np.arange(35).reshape(5,7)
# array([[ 0,  1,  2,  3,  4,  5,  6],
#        [ 7,  8,  9, 10, 11, 12, 13],
#        [14, 15, 16, 17, 18, 19, 20],
#        [21, 22, 23, 24, 25, 26, 27],
#        [28, 29, 30, 31, 32, 33, 34]])

and select the 2nd and 3rd elements of the 1st, 3rd and 5th array like so:

y[np.array([0,2,4]), 1:3]
# array([[ 1,  2],
#        [15, 16],
#        [29, 30]])

I cannot find a way to replicate this using arrays in place of the slice for indexing, the following doesn't work, I must be able to use arrays to index as I sometimes might be interested in the 2nd and 4th elements of the arrays and so on:

y[np.array([0,2,4]), np.array([1,2])]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
IndexError: shape mismatch: indexing arrays could not be broadcast together with shapes (3,) (2,)

How can I achieve my desired functionality?

You need to use a list, just as you did with the first dimension. — Prune
– Prune, Commented Apr 6, 2020 at 20:04
As you can see, that doesn't work, the indexing arrays could not be broadcast together — H. Wilde
– H. Wilde, Commented Apr 6, 2020 at 20:04
I see where you use a list and an array; I do not see where you've used two lists. — Prune
– Prune, Commented Apr 6, 2020 at 20:06
Do you mind giving an example of what you mean, I am not sure I follow? — H. Wilde
– H. Wilde, Commented Apr 6, 2020 at 20:07
broadcasting with indices is the same as with addition or multiplication. np.array([1,2])[:,None] + np.array([1,2,3]) produces a (2,3) sum. arr[np.array([1,2])[:,None], np.array([1,2,3])] indexes a (2,3) block. — hpaulj
– hpaulj, Commented Apr 6, 2020 at 20:33

Ethan · Accepted Answer · 2020-04-06 20:52:29Z

2

np.ix_() is designed for this type of problem.

def getSub():
    y = np.arange(35).reshape(5,7)
    locs = np.ix_([0,2,4],[1,2])
    return y[locs]


>>> getSub()
array([[ 1,  2],
       [15, 16],
       [29, 30]])

answered Apr 6, 2020 at 20:52

Ethan

1,3831 gold badge8 silver badges8 bronze badges

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

Comments

Amit Aviv · Accepted Answer · 2020-04-06 20:25:45Z

0

y[np.array([0,2,4]), np.array([[1],[2]])].T

answered Apr 6, 2020 at 20:25

Amit Aviv

1,81611 silver badges10 bronze badges

Comments

RDKM · Accepted Answer · 2020-04-06 20:39:09Z

0

you can try to use this y[np.array([[0,2,4]]*2),np.array([[1]*3,[2]*3])].T

answered Apr 6, 2020 at 20:39

RDKM

11 bronze badge

Comments

Axeon Thra · Accepted Answer · 2020-04-06 20:43:23Z

0

Quick, dirty way to achieve it with double indexing.

y[np.array([0,2,4]),:][:,np.array([1,2])]

answered Apr 6, 2020 at 20:43

Axeon Thra

3744 silver badges15 bronze badges

Collectives™ on Stack Overflow

Select using multiple arrays of indices from an ndarray in numpy

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related