How can I make the vector output always a column vector?

I'm having trouble with a problem where I need the output vector to be a column vector always. The output sometimes provided as row vectors and in the iteration stops as vector dimensions are not the same. The algorithm stops after sometime when the output takes row vector. Please help me how to fix the out of a vector always as a column vector. here is one example.

$x{n+1}=z{n}$ is not predictable. it gives sometimes as a row vector and sometimes as a column vector. it supposed to finish all 2000 iterations, but it will give me error sometimes 

N=100;
matrixSize=N;
x{1}=fix(randi([5,5],N,1));
W=fix(randi([2,2],N,1));
for n=1:200; 
y{n}=x{n}+W;
for j=1:N;
  if y{n}(j,:)<0;
      z{n}(j,:)=100;
  else
      z{n}(j,:)=99;
  end
  x{n+1}=z{n};
end
end

[Bruno Luong]

Your code has plenty of row vectors, and not column vectors.

a(j,:) % <- row
a(:,j) % <- column

Unless if you leave in a parallel universe than the rest of us.

Answer 1

[madhan ravi]

N=100;
matrixSize=N;
x{1}=fix(randi([5,5],N,1));
W=fix(randi([2,2],N,1));
for n=1:200; 
y{n}=x{n}'+W;
for j=1:N;
  if y{n}(j,:)<0;
      z{n}(j,:)=100;
  else
      z{n}(j,:)=99;
  end
  x{n+1}=z{n};
end
end

[madhan ravi]

problem solved!

[Dinoma Degefa]

Dear Sir, here is my code. but not working. after going a few iterations it will stop as the outcome for q4{n,:}=RR{n,:} is a row vector. run it reputedly you will see.

N=100;
matrixSize=N;
x{1,:}=fix(randi([50,50],N,1));
x{2,:}=fix(randi([200,200],N,1));
for n=2:2000; 
beta(n)=(1/3)-(1/(n+2));
WWW{n,:}=beta(n)*x{n,:}-beta(n)*x{n-1,:};
WW1{n,:}=x{n,:}+WWW{n,:};

AA=eye(N);
y{n,:}=AA*x{n,:};

R{n,:}=AA*WW1{n,:};

   for j=1:N;
          if norm(R{n,:}(j,:))<1;
              RR{n,:}(j,:)=R{n,:}(j,:);
          else if norm(R{n,:}(j,:))<=2 & norm(R{n,:}(j,:))>=1;
                  if R{n,:}(j,:)<0;
                      RR{n,:}(j,:)=-1;
                  else   if R{n,:}(j,:)>0;
                      RR{n,:}(j,:)=1;
                      else RR{n,:}(j,:)=0;
                      end
                  end
                  else   if R{n,:}(j,:)-1<0;
                      RR{n,:}(:,j)=-1;
                  else   if R{n,:}(j,:)-1>0;
                      RR{n,:}(j,:)=y{n,:}(j,:)-1;
                      else RR{n,:}(j,:)=0;

            end
                      end
              end
          end
   end

          q4{n,:}=RR{n,:}; 

      x{n+1,:}= y{n,:}-RR{n,:};       %(I-proxg2)Ax(n)
  end
  toc;