How to sort rows of a 2D array based on another array

Hi,
I would like to ask how can I sort rows of one 2D array based on values in another 2D array with same dimensions without using loop. For example: Sort rows of B based on rows of A
A = [1 2 3;2 1 3;3 2 1]
B = [0 1 2;0 1 2;0 1 2]
the result is newB = [0 1 2;1 0 2;2 1 0]
My gimmicky solution using loop is:
[~,idx] = sort(A,2)
for i=1:height(B)
B(i,:) = B(i,idx(i,:))
end
Is there cleaner slution?
Thanks.
Vojta

 Accepted Answer

Here's one example. I'm sure there are others.
A = [1 2 3;2 1 3;3 2 1];
B = [0 1 2;0 1 2;0 1 2];
% get sorting and size info
sz = size(B);
[~,cidx] = sort(A,2);
% sort in a loop
C = zeros(sz);
for r = 1:sz(1)
C(r,:) = B(r,cidx(r,:));
end
C
C = 3×3
0 1 2 1 0 2 2 1 0
% use sub2ind()
ridx = repmat((1:sz(1)).',[1 sz(2)]);
idx = sub2ind(sz(1:2),ridx,cidx);
D = B(idx)
D = 3×3
0 1 2 1 0 2 2 1 0

More Answers (1)

Zdravím kolego, // Hello,
I have tried something. I can't say it's "cleaner solution", but it's without for loop. You basically unfold matrix into one vector, and then you do not need a loop.
% original variables
A = [1 2 3;2 1 3;3 2 1];
B = [0 1 2;0 1 2;0 1 2];
% modified variables -- make them into a vector
newB = B';
newB = newB(1:numel(B));
newIdx = A';
newIdx = newIdx(1:numel(B));
final_B = newB(newIdx);
final_B = reshape(final_B,[3 3])'
final_B = 3×3
0 1 2 1 0 2 2 1 0

1 Comment

Dobrý den/Hi,
thanks for the idea. Unfolding the matrix into a vector makes it easier.

Sign in to comment.

Categories

Products

Release

R2022a

Tags

Community Treasure Hunt

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

Start Hunting!