Finding the non-intersecting rows in a matrix

Question

I want to find the non-intersecting rows in a large matrix. As an example:

A=[1 5 3; 3 4 5; 7 9 10;4 5 6;11 2 8; 3 5 10]

In this matrix, the non-intersecting rows are: [1 5 3], [11 2 8] and [7 9 10] . How can I program this in Matlab in a fast way?

Answer 1

If I may bsxfun -

M = squeeze(any(bsxfun(@eq,A,permute(unique(A),[3 2 1])),2))
[~,row_idx] = max(M,[],1)
out = A(sum(M,2).' == histc(row_idx,1:size(A,1)),:)

Sample step-by-step run -

A =
     1     5     3
     3     4     5
     7     9    10
     4     5     6
    11     2     8
     3     5    10
M =
     1     0     1     0     1     0     0     0     0     0     0
     0     0     1     1     1     0     0     0     0     0     0
     0     0     0     0     0     0     1     0     1     1     0
     0     0     0     1     1     1     0     0     0     0     0
     0     1     0     0     0     0     0     1     0     0     1
     0     0     1     0     1     0     0     0     0     1     0
row_idx =
     1     5     1     2     1     4     3     5     3     3     5
out =
     1     5     3
     7     9    10
    11     2     8

Answer 2

You can look for rows that adding them to union of previous rows increases the number of elements in the union by the number of columns (ie all elements in that row are new):

B = []; 
C = zeros(1,size(A,1)); 
for k=1:size(A,1), 
    B1 = union(B, A(k,:)); 
    C(k) = numel(B1)-numel(B); 
    B=B1; 
end
result = A(C==size(A,2),:);