How to sort minimum value from a matrix

Hi there, I have 2 set of array such as

number = 6;
A = 1+10*(rand(number,2));
B = [4,8; 6,8; 8,8; 6,6.7; 6,5.4; 6,4.1];

% Calculate all distance between set A and B for each elements
s = size(A);      
dist = cell(1,s(1));
for idx = 1:s(1)
    dist{idx} = vecnorm(B - A(idx,:),2,2); 
end    
dist = cell2mat(dist);
[minDist,index] = min(dist(:,:));   % obtain min distance and their index

The first set A is generated randomly and the second set B is fixed, and after running the code it will give a list of results such as

each column represent the distance from each element from set A to all in the set B. And the index indicate in which row the minium value located at. Hence, I was wondering how can sort the minimun value out, given that there might be huge posiballity of one of the point have the same minimun distance point as the figure show for index. If than happen, I would like to find the second minimum value to be the minimin value to ensure it wont obtain the same index of minimun value. For example, the index I would like to have the number 1 to 6 for each column and not repeated. Thanks.

-Chann-

[KSSV]

Read about knnsearch

Answer 1

[Bruno Luong]

See matchpairs function

number = 6;
A = 1+10*(rand(number,2));
B = [4,8; 6,8; 8,8; 6,6.7; 6,5.4; 6,4.1];
% Calculate all distance between set A and B for each elements
s = size(A);
dist = cell(1,s(1));
for idx = 1:s(1)
    dist{idx} = vecnorm(B - A(idx,:),2,2);
end
dist = cell2mat(dist)

dist = 6×6

2.0023 3.2585 6.6591 5.7172 5.5799 6.7287 1.3561 4.7442 5.4880 4.0241 4.9902 4.8789 2.7693 6.5113 4.8880 2.7749 5.1642 3.2142 2.6169 5.5625 4.3797 4.9478 3.7012 4.4545 3.9035 6.5387 3.4096 6.0123 2.4240 4.3889 5.1968 7.6124 2.7294 7.1551 1.1970 4.6970

M = matchpairs(dist,max(dist,[],'all'));
for k=1:size(M,1)
    fprintf('Column %d matches row %d, dist(i,j) = %f\n', M(k,2), M(k,1), dist(M(k,1),M(k,2)))
end

Column 1 matches row 2, dist(i,j) = 1.356077
Column 2 matches row 1, dist(i,j) = 3.258484
Column 3 matches row 5, dist(i,j) = 3.409639
Column 4 matches row 3, dist(i,j) = 2.774933
Column 5 matches row 6, dist(i,j) = 1.196998
Column 6 matches row 4, dist(i,j) = 4.454540

[Shin]

Hi Bruno, thank you very much!! appreciate the help.

Answer 2

[KSSV]

number = 6;
A = 1+10*(rand(number,2));
B = [4,8; 6,8; 8,8; 6,6.7; 6,5.4; 6,4.1];
idx = knnsearch(A,B)

idx = 6×1

6 2 4 2 3 3

[Shin]

Hi KSSV, thanks for your reply, however this method makes not different, the index will still appear to be repeating.

Answer 3

[Image Analyst]

Try this:

number = 6;
A = 1+10*(rand(number,2))
B = [4,8; 6,8; 8,8; 6,6.7; 6,5.4; 6,4.1]
distances = pdist2(A, B)
% Find unique distances because it's possible some may be repeated.
udistances = unique(distances) % It's also sorted from in to max
% Find row and column for all the distances, from shortest to longest in sorted order.
for k = 1 : numel(udistances)
    thisDistance = udistances(k);
    fprintf('Min distance of %f at\n', thisDistance);
    [indexA, indexB] = find(distances == thisDistance) % It's possible there will be multiple indexes if using integer coordinates.
end

[Shin]

Hi there, thanks for your solution.