error:index out of bounds

folks, can someone help me to figure out where is the mistake from this simple code. I'm getting error "Attempted to access data_matrix(:,169); index out of bounds because size(data_matrix)=[18,168]."

 for f=1:blockSizeR
    for g=1:blockSizeC

        Ism = X(blockVectorR:blockRR , blockVectorC:blockCC);
        data_matrix = [data_matrix,feat_vect];
end
end

ids_train = [];
  train_data = [];

    cont = 1;
    for ii=1:100
        for jj=1:6                                     
            if jj<2
                train_data = [train_data,data_matrix(:,cont)];
                ids_train  = [ids_train, ids(cont)];
             else
                test_data = [test_data,data_matrix(:,cont)];
                ids_test  = [ids_test,ids(cont)];
            end
            cont = cont + 1;
        end
    end

[Adam]

Are you sure that is the piece of code giving that error? I don't see anywhere in that code where you are indexing into data_matrix.

[Seprienna]

Actually this is the part where the error comes from Any advise?

ids_train = [];
     train_data = [];

    cont = 1;
    for ii=1:100
        for jj=1:6                                     
            if jj<2
                train_data = [train_data,data_matrix(:,cont)];
                ids_train  = [ids_train, ids(cont)];
                else
                test_data = [test_data,data_matrix(:,cont)];
                ids_test  = [ids_test,ids(cont)];
            end
            cont = cont + 1;

        end

    end

[A Jenkins]

What are you trying to do? cont is going to increase each time 6 times through the inner loop, then each time 100 times that through the outer loop, so it is going to count up to 601 by the time your loop is done. If your data_matrix is not that big, then what elements would you like to select from it?

Answer 1

[Adam]

Your 'cont' variable is ranging up to > 600 after 100 outer loops and 6 inner loops so if you only have a 2nd dimension of 168 in your data_matrix then that would be the reason.

Did you perhaps mean to increment cont inside the if statement? i.e.

ids_train = [];
train_data = [];
cont = 1;
for ii=1:100
    for jj=1:6                                     
        if jj<2
            train_data = [train_data,data_matrix(:,cont)];
            ids_train  = [ids_train, ids(cont)];
        else
            test_data = [test_data,data_matrix(:,cont)];
            ids_test  = [ids_test,ids(cont)];

              cont = cont + 1;
          end
      end
 end

That would keep 'cont' to a maximum value of 100 though whether that is what you want if you have 168 elements in that 2nd dimension I couldn't say.

[Seprienna]

maybe i didn't explain in details. cont is to calculating the total number which in this case i'm looking at 601. let me attached a fuller code here, if i add-in verbose,it seems to work where cont-1, but i end up getting a different error. "Assignment has more non-singleton rhs dimensions than non-singleton subscripts"

I'm not sure which part of the coding i shall change,please advise

if(verbose)
              disp(sprintf('Finished assigning image %i/%i to appropriate image set',cont-1,400));
          end

the error come from here instead.

 results.match_dist = zeros(d,b);

    for i=1:d
        for j=1:b
            results.match_dist(i,j) = return_PhD_distance(train(1:n,j),test(1:n,i),dist,covar);
            if train_ids(j)==test_ids(i)
                results.same_cli_id(i,j)=1;
            else
                results.same_cli_id(i,j)=0;
            end
        end
    end