indexing in numpy nd-array

Question

I have a numpy nd array with the following shape as an example:

import numpy as np

# Note this can be variable and have different dimensions each time
shape = [100, 100, 100, 5]
array = np.zeros(shape)

Now what I want to do is address the last dimension, so I need something like:

array[:, :, :, i] = something

But I cannot hard code it like this as the dimensions can change depending on the input. So is there a way to index the last dimension of an nd-array?

user2357112 · Accepted Answer · 2014-07-25 12:20:35Z

4

Yup:

array[..., i] = something

That's a literal ... there, not a placeholder for something else you're expected to write.

answered Jul 25, 2014 at 12:20

user2357112

286k32 gold badges490 silver badges569 bronze badges

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

2 Comments

Dan Over a year ago

did not know about this one, thanks. Also nice to know that array[...,i,:], and array[i,...] work as expected. If one had a large array (many dimensions), would there be a way to get at the nth index or nth to last index other than stacking :s at one end?

user2357112 Over a year ago

@Dan: The easiest way I know of would be to create a view of the array with the desired axis moved to the front using numpy.rollaxis, then index the new view. For example, to access index k in the nth axis, we'd use numpy.rollaxis(array, n)[k].

Collectives™ on Stack Overflow

indexing in numpy nd-array

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related