Using one array as a multi-dimensional index into another

Asked Jun 19 '22 at 04:50

Active Jun 19 '22 at 04:50

Viewed 49 times

Say I have two arrays, G and gamma.

>>> G
array([[1, 2, 0],
       [1, 1, 1]])
>>> gamma
array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]],

       [[19, 20, 21],
        [22, 23, 24],
        [25, 26, 27]]])

I would like to use each row in G as a multi-dimensional index into gamma, i.e. look up the values in gamma with G. More concretely, I want to get array([16, 14]), since

>>> gamma[1, 2, 0]
16
>>> gamma[1, 1, 1]
14

I tried things like gamma[G] and np.take(gamma, G), but nothing gives me the correct result.

asked Jun 19 '22 at 04:50

nalzok

14,965
21
72
139

1

You want something like this `gamma[ G[0][0] ][ G[0][1] ][ G[0][2] ]`? – Carl HR Jun 19 '22 at 04:57
1

think in terms of indexing with tuples, e.g. `gamma[(1, 2, 0)]` – hpaulj Jun 19 '22 at 05:02
@CarlHR Yes, except that `g.shape[1]` may not be 3. – nalzok Jun 19 '22 at 05:08
@hpaulj Thanks! The answer is `np.array([gamma[tuple(idx)] for idx in G])`. – nalzok Jun 19 '22 at 05:10
`gamma[tuple(g.T)]` is all you need. Look at the `tuple(g.T)` by itself. – hpaulj Jun 19 '22 at 15:36
@hpaulj Oh I like that. Can confirm it works! – nalzok Jun 19 '22 at 21:05

Using one array as a multi-dimensional index into another

0 Answers0