How to compare multiple arrays from dictionary values, and map per array element the dictionary key into new array/list

Question

I want to compare multiple arrays which are defined in the values of a dictionary. I want to compare every element within all arrays to find the lowest value for this element. The key of the array corresponding to the lowest element should be mapped into a new array (or list) as such:

img_dictionary = {0: np.array([[935, 925, 271, 770, 869, 293],
                               [125, 385, 291, 677, 770, 770]], dtype=uint64),
                  1: np.array([[ 12,  92,  28, 942, 124, 882],
                               [241, 853, 292, 532, 834, 231]], dtype=uint64),
                  2: np.array([[934, 633, 902, 912, 922, 812],
                               [152, 293, 284, 634, 823, 326]], dtype=uint64),
                  3: np.array([[362,  11, 292,  48,  92, 481],
                               [196, 294, 446, 591,  92, 591]], dtype=uint64)}

Expected outcome

(array([[1, 3, 1, 3, 3, 0],
        [0, 2, 2, 1, 3, 1]], dtype=uint64),)

I have tried to use np.minimum() and functions alike, but these do not allow for the comparison of multiple arrays.

Ivan, you edited my np.array and np.dtype to array and dtype respectively, but if I try this without defining np. it states NameError: name 'array' is not defined, same for dtype NameError: name 'uint64' is not defined.. — Rivered
– Rivered, Commented Aug 4, 2021 at 17:20
Sorry about that, it was meant to be np.array. It will work as array if you import NumPy with from numpy import *... — Ivan
– Ivan, Commented Aug 4, 2021 at 18:44

Ivan · Accepted Answer · 2021-08-04 18:41:11Z

4

Assuming each entry in your dictionnary has the same number of elements, then you can work on your data as one array. This can be done by stacking the arrays with np.stack:

>>> arr = np.stack(list(img_dictionary.values()))
array([[[935, 925, 271, 770, 869, 293],
        [125, 385, 291, 677, 770, 770]],

       [[ 12,  92,  28, 942, 124, 882],
        [241, 853, 292, 532, 834, 231]],

       [[934, 633, 902, 912, 922, 812],
        [152, 293, 284, 634, 823, 326]],

       [[362,  11, 292,  48,  92, 481],
        [196, 294, 446, 591,  92, 591]]], dtype=uint64)

Then, you can apply np.argmin:

>>> arr.argmin(axis=0)
array([[1, 3, 1, 3, 3, 0],
       [0, 2, 2, 1, 3, 1]])

edited Aug 4, 2021 at 18:41

answered Aug 4, 2021 at 16:35

Ivan

41.3k9 gold badges78 silver badges120 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Rivered Over a year ago

Thanks, please add arr to np,argmin(axis=0) --> np.argmin(arr, axis=0)

Collectives™ on Stack Overflow

How to compare multiple arrays from dictionary values, and map per array element the dictionary key into new array/list

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related