Let’s say I have a list with multiple arrays of different length:
L = [
array([-10, -8, -3, 2, 1, 5, 12]),
array([-9, -4, -1, 3, 5]),
array([-11, -5, -4, 0, 1, 5, 7, 13, 18, 22])
]
How can I find the index of the lowest value the most efficiently?
For my example, the minimum value is -11 and the index is (2, 0), so the output should be (0, 2).
>Solution :
A possible solution, which needs preprocessing the list of arrays with padding, because of the unequal lengths of the arrays in L:
L = [
np.array([-10, -8, -3, 2, 1, 5, 12]),
np.array([-9, -4, -1, 3, 5]),
np.array([-11, -5, -4, 0, 1, 5, 7, 13, 18, 22])
]
max_size = np.max([x.size for x in L])
max_value = np.max([np.max(x) for x in L])
L_padded = [np.pad(x, (0, max_size - x.size),
constant_values= 1+max_value) for x in L]
L_2d = np.array(L_padded)
min_index = np.unravel_index(L_2d.argmin(), L_2d.shape)
(min_index[0], min_index[1])