access multiple cell array values and compare with a matrix

%no of steps
nsteps=30;
%no of species
ncell=3;

for c=1:ncell
    cell{c}=c;
end

[B C D]=deal(cell{:});

I have generated matrix containing values mentioned above

A=zeros(ncell,nsteps);
A(B,1)=0;
A(C,1)=1;
A(D,1)=0;

I have a input matrix (a logic table) in text format IL1.txt

B  C  D
0  0  0
0  1  1
1  0  0
1  1  0

IL1_data = importdata('IL1.txt','\t');

   for  t=1:nsteps-1
       for n=1:size(IL1_data.data,1)
            if ((IL1_data.data(n,1:(end-1)))==[A(cell2mat(IL1_data.textdata(1,1:2)),t)])
                A(cell2mat(IL1_data.textdata(1,3)),t+1)=(IL1_data.data(n,end));
                break
            end
       end
  end

it should work like:

for  t=1:nsteps-1
   for n=1:size(IL1_data.data,1)
        if ((IL1_data.data(n,1:(end-1)))==[A(B,t) A(C,t)])
            A(D,t+1)=(IL1_data.data(n,end));
            break
        end
   end        
end

But it is giving an error: Index exceeds matrix dimensions

Thank you.

[Guillaume]

Calling a variable cell is a very bad idea since it shadows the cell function for creating cell arrays.

Answer 1

[Guillaume]

the problem comes from this:

A(cell2mat(IL1_data.textdata(1,1:2)),t)

which is equivalent to

A('BC', t)

If you pass a string as an index it is converted into its numerical value which is the ASCII value of the characters, so really what you're asking is:

A([66 67], t)

I'm not clear exactly on what you're trying to do with the variables B, C and D and the string headers. If theses variables are always 1, 2, and 3 respectively, one workaround would be to do:

indices = cell2mat(IL1_data.textdata(1, 1:2)) - 'A';
A(indices, t)

If what you really want is that string to refer to the actual variables B and C, then you could use eval, but I would not recommend going down that route:

indices = eval(sprintf('[%c %c]', IL1_data.textdata{1, 1:2}))

[Pooja Dnyane]

Thank you very much for your reply.

I am numbering B,C and D such that A(B,1) is A(1,1) and trying to use string header (same variable) to access values from matrix A. So, every time with different table you don't have to provide input manually.

what if we use Bcell Ccell Dcell or variables with other names instead of using B,C and D

[Guillaume]

If the rows in A are always going to be in the order 'B', 'C', 'D', etc. then my first workaround would be the way to go (subtract the ascii value of 'A' from the name of the columns in your text file.

Another fairly elegant solution (but with some overhead) is to use tables instead:

nsteps = 30;
ncells = 3;
A = array2table(zeros(ncells, nsteps), 'RowNames', {'B', 'C', 'D'}, ...
    'VariableNames', arrayfun(@(s) sprintf('step_%02d', s), 1:nsteps, 'UniformOutput', false));

A{{'B', 'C', 'D'}, 1} = [0 1 0]';

IL1_data = importdata('IL1.txt','\t');
for t = 1:nsteps
    matchingrow = ismember(IL1_data.data(:, 1:2), A{IL1_data.textdata(1:2), t}', 'rows');
    if sum(matchingrow) == 1 %found one row that match
       A{IL1_data.textdata(3), t+1} = IL1_data.data(matchingrow, 3);
    end
end

Note that the code above does exactly the same as your pseudocode, that is put in the next column the value of 'D' corresponding to the 'B' and 'C' of the current column. As it never touches rows 'B' and 'C', I don't think it's the result you want.