Numpy: Assignment and Indexing as Matlab

Question

Sometimes is useful to assign arrays with one index only. In Matlab this is straightforward:

M = zeros(4);
M(1:5:end) = 1
M =

   1   0   0   0
   0   1   0          


        

Answer 1

You could try numpy.ndarray.flat, which represents an iterator that you can use for reading and writing into the array.

>>> M = zeros((4,4))
>>> M.flat[::5] = 1
>>> print(M)
array([[ 1.,  0.,  0.,  0.],
       [ 0.,  1.,  0.,  0.],
       [ 0.,  0.,  1.,  0.],
       [ 0.,  0.,  0.,  1.]])

Note that in numpy the slicing syntax is [start:stop_exclusive:step], as opposed to Matlab's (start:step:stop_inclusive).

Based on sebergs comment it might be important to point out that Matlab stores matrices in column major, while numpy arrays are row major by default.

>>> M = zeros((4,4))
>>> M.flat[:4] = 1
>>> print(M)
array([[ 1.,  1.,  1.,  1.],
       [ 0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.]])

To get Matlab-like indexing on the flattened array you will need to flatten the transposed array:

>>> M = zeros((4,4))
>>> M.T.flat[:4] = 1
>>> print(M)
array([[ 1.,  0.,  0.,  0.],
       [ 1.,  0.,  0.,  0.],
       [ 1.,  0.,  0.,  0.],
       [ 1.,  0.,  0.,  0.]])