find 2 common values of 3 in row matrix?

HI,

I have a 3 columns and n rows matrix I need to find in this matrix, witch rows own the same 2 values of the 3 as the first row. the values can be in a other column order. then find witch rows own the same 2 values of the 3 as the second row...etc

example: A=[1,2,3;3,2,4;5,6,7;7,6,8]

the first loop must find that row 1 has 2 commun values with row 2. third loop must find that row 3 has 2 commun values with row 4.

[Image Analyst]

Are these guaranteed to be integers , or can they be floating point numbers?

Answer 1

[dpb]

>> for i=1:2:size(A,1),cmn=intersect(A(i,:),A(i+1,:)),end
cmn =
   2     3
cmn =
   6     7
>>

[Stephen23]

Note that i and j should not be used for variable names, as these are both names of the inbuilt imaginary unit .

Answer 2

[nicolas]

I have [written] this code, it works, but it is very slow, because of the two if loops. Is it possible to optimized it?

for i=1:size(Lien,1)
  k=1;
  for j=1:size(Lien,1)
    sommetscommuns=intersect(Lien(i,:),Lien(j,:))
    if size(sommetscommuns,2)==2
      voisins(i,k)=j;
      k=k+1;
    end
  end
  if voisins(i,1) ~=0 && voisins(i,2) ~=0 && voisins(i,3) ~=0
    Qnor(i,1)=(dot(Normales(i,:),Normales(voisins(i,1),:)) + ...
               dot(Normales(i,:),Normales(voisins(i,2),:)) + ...
               dot(Normales(i,:),Normales(voisins(i,3),:)))/3;
  else
    Qnor(i,1)=10;
  end
end

th[a]nks of lot

[dpb]

1) Have you preallocated Qnor? If not, that'll help noticeably if sizes are very large

2) Is this not symmetric in the end? If so, running the loop only over the triangular section would cut it roughly in half. Also, the i==j locations are all identities.

Just one minor syntax that won't have any significant impact on speed would be

if voisins(i,1) ~=0 && voisins(i,2) ~=0 && voisins(i,3) ~=0

can be written as

if all(voisins(i,:))