Basics of numpy where function, what does it do to the array?

Question

I have seen the post Difference between nonzero(a), where(a) and argwhere(a). When to use which? and I don\'t really understand the use of the where function from numpy modu

Answer 1

np.where returns indices where a given condition is met. In your case, you're asking for the indices where the value in Z is not 0 (e.g. Python considers any non-0 value as True). Which for Z results in:

(0, 0) # top left
(0, 2) # third element in the first row
(0, 3) # fourth element in the first row
(1, 3) # fourth element in the second row
...    # and so on

np.where starts to make sense in the following scenarios:

a = np.arange(10)
np.where(a > 5) # give me all indices where the value of a is bigger than 5
# a > 5 is a boolean mask like [False, False, ..., True, True, True]
# (array([6, 7, 8, 9], dtype=int64),)

Hope that helps.