Indexing multidimensional numpy-array with list

Question

I have a problem accessing the data in a multidimensional numpy-array with python 2.7. The goal is to read multiple values whose positions are stored in a list.

import numpy as np
matrix=np.ones((10,30,5))

positions=[]
positions.append([1,2,3])
positions.append([4,5,6])

for i in positions:
    print matrix[i]

What I want is:

print matrix[1,2,3]

But I get:

print [matrix[1], matrix[2], matrix[3]]

Could you please give me a hint for the correct indexing? Thanks!

Answer 1

From indexing docs :

In Python, x[(exp1, exp2, ..., expN)] is equivalent to x[exp1, exp2,..., expN]; the latter is just syntactic sugar for the former.

So, instead of passing it a list, pass a tuple to matrix :

for i in positions:
    print matrix[tuple(i)]

Lists are used for picking item at particular indices, ie index arrays :

>>> arr = np.random.rand(10)
>>> arr
array([ 0.56854322,  0.21189256,  0.72516831,  0.85751778,  0.29589961,
        0.90989207,  0.26840669,  0.02999548,  0.65572606,  0.49436744])
>>> arr[[0, 0, 5, 1, 5]]
array([ 0.56854322,  0.56854322,  0.90989207,  0.21189256,  0.90989207])