Strange bug when indexing vector

1 view (last 30 days)
Matt Flood
Matt Flood on 14 Dec 2021
Commented: Matt Flood on 14 Dec 2021
The following piece of code works as I want it to.
It sorts the elements in a 2x2 matrix and gives equal rank order to elements with the same value.
x = [2.3333 2.3333; 2.0000 2.3333];
[y, z] = sort(x(:))
y = 4×1
2.0000 2.3333 2.3333 2.3333
z = 4×1
2 1 3 4
if any(diff(y)==0)
for n = find(diff(y')==0)+1
z(n) = z(n-1);
end
end
disp(z)
2 1 1 1
So x is sorted as [2, 2.33, 2.33, 2.33] and the elements are given the ranks [2, 1, 1, 1].
However, if I don't transpose y in the for loop, the code gives the wrong output, i.e., the rank is [2, 1, 1, 3].
x = [2.3333 2.3333; 2.0000 2.3333];
[y, z] = sort(x(:))
y = 4×1
2.0000 2.3333 2.3333 2.3333
z = 4×1
2 1 3 4
if any(diff(y)==0)
for n = find(diff(y)==0)+1
z(n) = z(n-1);
end
end
disp(z)
2 1 1 3
As far as I can tell, there is no logical reason why not transposing y vector should produce a different output.
Can someone please explain why this is happening??
  2 Comments
Voss
Voss on 14 Dec 2021
Edited: Voss on 14 Dec 2021
for loops in MATLAB loop over the columns of the variable you tell it. So, for example, this:
for i = [1 2 3]
display(i);
end
i = 1
i = 2
i = 3
is as expected (columns of a row vector are scalars), but the following only loops once and i is the whole column vector:
for i = [1; 2; 3]
display(i);
end
i = 3×1
1 2 3
Generalizing to a 2d matrix:
x = magic(3);
for i = x
display(i);
end
i = 3×1
8 3 4
i = 3×1
1 5 9
i = 3×1
6 7 2
So in your second example (without transposing y), n takes the value [3; 4] and the loop iterates once, rather than the loop iterating twice with n taking 3 and then 4, which is what it does in the first example.
Matt Flood
Matt Flood on 14 Dec 2021
Thank you @Benjamin
I see it now - it's treating [3;4] as a single interable as opposed to two individual ones.

Sign in to comment.

Answers (0)

Categories

Find more on Sparse Matrices in Help Center and File Exchange

Products


Release

R2020b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!