Store unique values ​​in an array based on another array while maintaining order

Question

I have two arrays my_arrand distances. For instance:

my_arr= np.array([0, 1, 3, 4, 5, 3, 5])
distances = np.array([18, 47, 20, 10, 26, 22, 13])

I would like to get an array of indices whose shape is np.unique(my_arr).sizebased on the minimum distance. So in the previous example, I would get:

# indices of my_arr
indices_of_my_arr= np.array([0, 1, 2, 3, 6])

With the exception of forloops, or mapis there a tricky way to do this?

EDIT: Another example:

my_arr = np.array([0, 2, 3, 1, 3, 4, 4, 5])
dist = np.array([10, 12, 15, 18, 5, 14, 45, 8])

I expect:

[0, 1, 3, 4, 5, 7]

Answer 1

You can use np.lexsortand np.unique-

idx = np.lexsort([distances, my_arr])
out = np.sort(idx[np.unique(my_arr[idx], return_index=1)[1]])