Convert max numpy array offset to a tuple?

Question

Is it possible to convert the largest numpy array offset to a tuple?

So for example, if "array" was:

[[1 2 3 6 8],
 [2 4 1 1 0],
 [0 0 0 20 0]]

Then the np.max(array) would return 20 and it's position/ offset is array[2][3].

Is it possible to turn array[2][3] into tuple = (2,3)?

Thank you for the input.

https://stackoverflow.com/questions/3584243/get-the-position-of-the-biggest-item-in-a-numpy-array — Khalil Al Hooti, Oct 31 '18 at 17:05

score 3 · Accepted Answer · answered Oct 31 '18 at 17:03

3

you are looking for unravel_index and argmax

import numpy as np

a = np.array([[1 2 3 6 8],
              [2 4 1 1 0],
              [0 0 0 20 0]])

np.unravel_index( a.argmax() , a.shape)

answered Oct 31 '18 at 17:03

Gabriel M

score 1 · Answer 2 · answered Oct 31 '18 at 17:03

1

To find all max value indices:

np.argwhere(np.max(a) == a)
# array([[2, 3]])

Then you can get the first max value index:

np.argwhere(np.max(a) == a)[0]
# array([2, 3])

And convert it to tuple if needed:

tuple(np.argwhere(np.max(a) == a)[0])
# (2, 3)

answered Oct 31 '18 at 17:03

Psidom

2 Answers2