Assigning values to specific subscript indices in 3D python array throws errors

Question

Say I have initialized an array in 3D as:

arr_3d = np.zeros((100, 100, 100))

Now I want to change the elements at subscript indices (i, j, k) of arr_3d to some value (say 1), where i, j, k are the list (or array) of indices with sizes 100, 100, 40, respectively, along the three axes. I have tried arr_3d[i, j, k] = 1, but it throws an error. I have tried converting the subscript indices to linear indices by np.ravel_multi_index(), but it looks like it can't convert the subscript indices of a 3D array.

The above issue is easy to solve in Matlab, where using arr_3d(i, j, k) = 1 works.

Answer 1

Try it with different dimensions,ones that are distinct

In [1375]: a=np.zeros((2, 3, 4))
In [1376]: a[np.ix_([0], [1,2], [0,1])] = 1
In [1377]: a
Out[1377]: 
array([[[ 0.,  0.,  0.,  0.],
        [ 1.,  1.,  0.,  0.],
        [ 1.,  1.,  0.,  0.]],

       [[ 0.,  0.,  0.,  0.],
        [ 0.,  0.,  0.,  0.],
        [ 0.,  0.,  0.,  0.]]])

numpy divides the array in to 'planes' on the first dimension; MATLAB does it on the last. So this shows 2 planes or blocks, each 3x4.

In octave

>> a = zeros(2,3,4);
>> a([1],[2,3],[1,2]) = 1
a =

ans(:,:,1) =

   0   1   1
   0   0   0

ans(:,:,2) =

   0   1   1
   0   0   0

ans(:,:,3) =

   0   0   0
   0   0   0

ans(:,:,4) =

   0   0   0
   0   0   0

4 blocks of 2x3.