Storing Outputs from a Nested Loop

I have the following problem, wherein, i want to store outputs from the nested loops. I am just presenting a simple scenario to depict the situation:

for a=[1:1:10]
  for b=[1:1:10]
    c=a*b
  end
end

c in this case would store the last value i.e. 100. How to develop an array which stores all the values of c from 1 to 100 as c(1,1)=1, c(1,2)=2 and so on.

Answer 1

[Iddo Weiner]

By initiating an empty array and then filling it with the looping indices:

c = nan(10,10);
for a = 1:10
  for b= 1:10
    c(a,b)=a*b;
  end
end
% visualize c
imagesc(c)

[Stephen23]

Note that the MATLAB documentation calls this preallocation:

https://www.mathworks.com/help/matlab/matlab_prog/preallocating-arrays.html

https://www.mathworks.com/help/matlab/matlab_prog/preallocate-memory-for-a-cell-array.html

[A M]

Thank you for this support. I am still stuck. Can you look into this code:

c=nan(5,5);
for a=[20:20:100];
  for b =[5:-1:1];
    c(a,b)=(a+b);
  end
end

Please correct me. To me belief, this should have returned a 5x5 matrix with all values stored in it such that the result of c(20,5)=25 is appeared. Instead it shows that 'c' is a 5x100 with 0 stored in 1-19 and then values appearing at 20. How to solve this issue so i can have a clean matrix of 5x5.

[Stephen23]

A = 20:20:100;
B = 5:-1:1;
C = NaN(numel(A),numel(B));
for a = 1:numel(A)
    for b = 1:numel(B)
        C(a,b) = A(a)+ B(b);
    end
end

You were using the values as indices. Instead you need to create separate indices and use those to index into the vectors and output matrix.

[Jan]

@A M: Do you see, that the suggested codes contain less square brackets? See http://www.mathworks.com/matlabcentral/answers/35676-why-not-use-square-brackets .

[A M]

Thank you. This appears to be a promising solution. It worked out and i got a 5*5 matrix. So, if i am not mistaken, the NaN creates a n*n matrix. What does the 'Numel' exactly does? Is it not the same as NaN(5,5) creating a 5*5 matrix?

[Jan]

@A M: See

doc numel

[Niels]

Matlab uses (as you should also do) smart shortcuts for its functions and variables, such as numel for Number of Elements

[Iddo Weiner]

I think you are supposed to get a 10*10 matrix in the example you posted. Yes, nan() initiates an empty array. You can also use zeros() to initiate a matrix of zeros. BTW - this code works also without the imitation line, i.e.:

for a = 1:10
  for b= 1:10
    c(a,b)=a*b;
  end
end

BUT this is highly un-recommended, because the shape of c changes constantly - which is a very inefficient use of memory. When you pre-allocate then c's shape is constant and the whole code runs much quicker (of course in the example case the array is very small and there won't be a significant difference).

Hope this helps, good luck

Answer 2

[Niels]

some changes:

aMax=10; % equal to #of rows
bMax=10; % equal to #of colums
c=zeros(aMax,bMax);
for a=1:aMax % let a run from 1 to aMax, stepwidth:=1
  for b=1:bMax
    c(a,b)=a*b; % hope you know that c is not running fom 1 to 100
  end
end

if you want c to cointain to values 1:(aMax*bMax) (100 in current state) change

c(a,b)=(a-1)*aMax+b;