Numpy create index/slicing programmatically from array

Question

I can use numpy.mgrid as follows:

a = numpy.mgrid[x0:x1, y0:y1] # 2 dimensional
b = numpy.mgrid[x0:x1, y0:y1, z0:z1] # 3 dimensional

Now, I'd like to create the expression in brackets programmatically, because I do not know whether I have 1, 2, 3 or more dimensions. I'm looking for something like:

shape = np.array([[x0, x1], [y0, y1], ... maybe more dimensions ...])
idx = (s[0]:s[1] for s in shape)
a = numpy.mgrid[idx]

That gives at least a syntax error in the second line. How can I properly generate those indices/slices programmatically? (The mgrid here is rather an example/use case, the question is really about indexing in general.)

Generate the ranges with idx = [range(s[0],s[1]) for s in shape] and then use meshgrid : np.meshgrid(*idx)? — Divakar
– Divakar, Commented Sep 20, 2016 at 20:54

wflynny · Accepted Answer · 2016-09-20 20:55:38Z

12

Use the slice object. For example:

shape = np.array([[0, 10], [0, 10]])
idx = tuple(slice(s[0],s[1], 1) for s in shape)
#yields the following
#(slice(0, 10, 1), slice(0, 10, 1))
np.mgrid[idx]

yields

array([[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
        [1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
        [2, 2, 2, 2, 2, 2, 2, 2, 2, 2],
        [3, 3, 3, 3, 3, 3, 3, 3, 3, 3],
        [4, 4, 4, 4, 4, 4, 4, 4, 4, 4],
        [5, 5, 5, 5, 5, 5, 5, 5, 5, 5],
        [6, 6, 6, 6, 6, 6, 6, 6, 6, 6],
        [7, 7, 7, 7, 7, 7, 7, 7, 7, 7],
        [8, 8, 8, 8, 8, 8, 8, 8, 8, 8],
        [9, 9, 9, 9, 9, 9, 9, 9, 9, 9]],

       [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]]])

Alternatively, you could use the Numpy shorthand np.s_, e.g. np.s_[0:10:1], instead of slice(1, 10, 1), but they are equivalent objects.

answered Sep 20, 2016 at 20:55

wflynny

18.6k6 gold badges50 silver badges69 bronze badges

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

3 Comments

Eric Over a year ago

slice(s[0],s[1], None) is a better translation of the intent

wflynny Over a year ago

Or even better slice(s[0], s[1]). I figured explicit is better than implicit for explanation.

hpaulj Over a year ago

Or with unpacking, [slice(s,e) for s,e in shapes]

Collectives™ on Stack Overflow

Numpy create index/slicing programmatically from array

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related