Is it possible to use cell indices (specified in large 3D arrays) to call specific data out of another 3D array?

I need to randomly sample data (70 samples, repeated 10,000 times) from a large temperature dataset (3D array) using cell indices specified in separate 3D arrays. I am able to generate 3D arrays specifying the cell indices for each dimension, but I am having issues using those indices to then call specific data from my original data set. When I try to use a for loop to sort through each dimension of the array and select the data I get the "subscript indices must either be real positive integers or logicals" which I realize has something to do with how I'm calling the data. My attempt to select data is at the end of my code below.

n = 70;
mc = 10000;

for i=1:9

      %randomly sample a depth using normal dist centered at m with stdev s
      mu = median(surf1:surf2);
      sigma = 1;

      depthr(:,:,i) = round(sigma .* randn(n,mc) + mu);
      depthr(depthr < surf1) = surf1;
      depthr(depthr > surf2) = surf2;

      %randomly sample cell numbers
      cellr(:,:,i) = round(length(tP) .* rand(n,mc));

      %select site data
      datar(:,:,i) = tP(:,:,min(find(round(lat) == round(corelat(i)))),...
          min(find(round(lon) == round(corelon(i)))));

end

for i=1:9
    for c=1:mc
        for r=1:n
            cell = cellr(r,c,i);
            depth = depthr(r,c,i);
            simr(r,c,i) = datar(cell,depth,i);
        end
    end
end

[Adam Danz]

Which line is producing the error? The one that produces 'datar' ?

[Brigitta Rongstad]

simr(r,c,i) = datar(cell,depth,i) line

Answer 1

[Adam Danz]

Without having the values to your variables, I can only guess.

Test these ideas

• the value of 'cell' is not a positive integer
• the value of 'depth' is not a positive integer

Answer 2

[Walter Roberson]

      cellr(:,:,i) = round(length(tP) .* rand(n,mc));

can be 0.

Perhaps it would make more sense to use randi() ?

Answer 3

[Brigitta Rongstad]

I realized one line of my code didn't exclude zeros from being randomly sampled. Once that is fixed, the code runs just fine!