Where did i mess those "for-loops" up?

" matrixg" is a 3d array, which contains incidence matrices of Graphs with 7 edges. This program should eliminate all Graphs which have a circle with 3 vertices.

I included a counter "g" and because of it, i think he realizes isomorphy as often as i think it should (19) times. Unfortunately the 3d-matrix H at then end still has size 33.

I think i either closed on "for" too early. But somehow i can't fix it myself.

clc

D=sparse([1 1 2],[2 3 3],true,3,3);

load('matrixg.mat')

c=[0 0 0 0 0 0 0];

a=1;

g=0;

[~, ~, z]=size(G);

for l=1:z

zaehler=0;

for i=1:7

%T1=[G(:,:,l) ; c];

T1=sparse(G(:,:,l));

T1(:,i)=[];

T1(i,:)=[];

for k=1:6

    T2=T1;

    T2(:,k)=[];

    T2(k,:)=[];

for l=1:5

    T3=T2;

    T3(:,l)=[];

    T3(l,:)=[];

for m=1:4

    T=T3;

    T(:,m)=[];

    T(m,:)=[];

[F1] = graphisomorphism(T+T',D+D','directed', false);

if F1==1

    zaehler=1; 

    g=g+1;

end

end

if zaehler==1

    break

end

    end

if zaehler==1

    break

end

end

if zaehler==1

    break

end

end

if zaehler==0

    H(:,:,a)=G(:,:,l);

    %view(biograph(H(:,:,a)));

    a=a+1;

end

end

[dpb]

a=1;
g=0;
for l=1:z
zaehler=0;
for i=1:7
  ...
  for k=1:6
    ...
    for l=1:5
      ...
      for m=1:4
        ....
        [F1] = graphisomorphism(T+T',D+D','directed', false);
        if F1==1
          zaehler=1; 
          g=g+1;
          break        % LOOKS TO ME LIKELY NEED THIS....
        end
      end
      if zaehler==1, break, end
    end
    if zaehler==1, break, end
  end
  if zaehler==1, break, end
end
if zaehler==0
  H(:,:,a)=G(:,:,l);
  a=a+1;
end
end

I recast your code a little including indenting uniformly and replacing the multi-line 'break' clauses w/ single to make flow easier to follow (imo such are better w/o the additional nesting).

It appears to me you likely need the additional break when first find the case in the innermost loop.

Otherwise, set up a dataset that has only a couple cases and use the debugger to step through and see where it isn't branching as you think should...