How to use find when some of the searches will not return a result?

I have some grid data stored in Excel file and I am trying to reformat them into grid data. Below is my code:

A = xlsread('Name.xlsx', 'Feb'); 
% 3 column data of Latitutde, Longitude, and Variable.
[X, Y] = meshgrid([2.5:5:357.5],[-76:4:80]);

[m n] = size(X);
for i=1:m;
    for j=1:n;
        if A( A(:,1)==Y(i,j), A(:,2)==X(i,j), 3) > 0;
            Z(i,j) = A( A(:,1)==Y(i,j), A(:,2)==X(i,j), 3);
        else
            Z(i,j) = NaN;
        end;
    end
end;

I am getting an "Index exceeds matrix dimensions" error. Clearly, some of the search are not supposed to retrieve results. How do I program it, so that it will automatically assign an NaN value, when index exceeds matrix dimensions?

Thanks.

Answer 1

[OCDER]

A is a Mx3 matrix, but you are trying to access A as a MxNx3 matrix, ex:

   A( A(:,1)==Y(i,j), A(:,2)==X(i,j), 3)  %A is not a 3D matrix, exceed dimension

Did you want A(C1&C2, 3), where C1 and C2 are 2 conditions that must be met? See code below with comments:

A = xlsread('Name.xlsx', 'Feb'); 

% 3 column data of Latitutde, Longitude, and Variable.
[X, Y] = meshgrid([2.5:5:357.5],[-76:4:80]);
Z = NaN(size(X)); %Preallocate Z, otherwise it gets REALLY slow
[m, n] = size(X);
for i=1:m %; Don't need semicolon
    for j=1:n %;
        C1 = A(:,1)==Y(i,j);    %pulling this out for readability and to prevent doing this 2 times
        C2 = A(:,2)==X(i,j);
        C  = C1 & C2;           %is this what you want? Must match conditions C1 and C2
        if A(C, 3) > 0          %FIX ERROR HERE?
            Z(i,j) = A(C, 3);   %Is this what  you want? 
%       else %Don't need this anymore, since  you initialized Z as NaN
%             Z(i,j) = NaN;
        end %;
    end
end %;

[Leon]

Thank you.

Answer 2

[Andrei Bobrov]

[X, Y] = meshgrid([2.5:5:357.5],[-76:4:80]);
[m n] = size(X);
Z = nan(m,n);
for ii=1:m
    for jj=1:n
        k = A( A(:,1)==Y(ii,jj) & A(:,2)==X(ii,jj), 3);
        if k > 0;
            Z(ii,jj) = k;
        end
    end
end

or

aa = reshape(A,[3,2,1]);
l0 = bsxfun(@eq,Y,aa(:,1,:)) & bsxfun(@eq,X,aa(:,2,:));
l1 = any(l0,3);
Z = sum(bsxfun(@times,l0,aa(:,3:)),3);
Z(~l1) = nan;

[Leon]

Many thanks for the codes and the multiple approaches! They work very well. I wish I could accept two answers.