Elementwise operation with different dimensions of vectors

General Usage
1

I am starting to write some codes in Julia and I have some doubts about elementwise operations.

Imagine that I have two matrices, A and B:

A= randn(3,3,3);
B = rand(3,3);

When I do something like A .==B what the compiler is doing is comparing

A[:,:1] .== B and A[:,:,2] .==B and A[:,:,3] .== B

and returning a boolean matrix (3,3,3) with the result? Is this correct?
The same applies for elementwise multiplication?
Is this a bad practice or leads to slow performance ? Is there a better way of doing this?

Thanks!

2

What do you mean? It is definitely intended to behave this way and meant to be used :slight_smile:
Do you mean that you do not want the whole array? In which case:

A = randn(3,3,3)
B = A[:,:,1]

compare_array(A, B) = A .== B

compare_bool(A, B) = all(k -> view(A, :, :, k) == B, axes(A, 3))

using BenchmarkTools
@btime compare_array($A, $B);
@btime compare_bool($A, $B);

compare_array is your function. As you said it returns the whole array and tells you precisely if there was a match or not and where

compare_bool performs the same tests but returns true or false such that all(compare_array(A, B)) == compare_bool(A, B).