Iterate over numpy matrix of unknown dimension

Question

I have a multidimensional numpy array I'd like to iterate over. I want to be able to access not only the values, but also their indices. Unfortunately,

for idx,val in enumerate(my_array):

doesn't seem to work when my_array is multidimensional. (I'd like idx to be a tuple). Nested for loops might work, but I don't know the number of dimensions of the array until runtime, and I know it's not appropriate for python anyway. I can think of a number of ways to do this (recursion, liberal use of the % operator), but none of these seem very 'python-esque'. Is there a simple way?

`enumerate` always returns an integer for the first value. What specifically are you expecting idx to contain? — ashastral, Jul 28 '12 at 00:48

score 10 · Accepted Answer · answered Jul 28 '12 at 00:49

10

I think you want ndenumerate:

>>> import numpy
>>> a = numpy.arange(6).reshape(1,2,3)
>>> a
array([[[0, 1, 2],
        [3, 4, 5]]])
>>> list(numpy.ndenumerate(a))
[((0, 0, 0), 0), ((0, 0, 1), 1), ((0, 0, 2), 2), ((0, 1, 0), 3), ((0, 1, 1), 4), ((0, 1, 2), 5)]

answered Jul 28 '12 at 00:49

DSM

342,061
65
592
494

argv! I just looked up the documentation to post a link. – mgilson Jul 28 '12 at 00:49
@mgilson: hey, now that you're a member of the 10k crew you can see how often that happens to all of us. ;^) – DSM Jul 28 '12 at 00:50

Iterate over numpy matrix of unknown dimension

1 Answers1

Linked