I need to work with variables at each iteration of a loop in Matlab whose names depend on the loop index h (e.g. if h=1 I want to use data1 e to create other variables). Is there a way to do it? I cannot use cells, because the variables are very large matrices and I have memory problems using cells.
Example:
data1=[1,2,3];
data2=[4,5,6];
data3=[7,8,9]; %they are in the workspace
for h=1:3
% A`h'=data`h'+6
% save A`h'
end
I think you should consider using structure with dynamic field names (see more details here).
For example
for h=1:n
dataName = sprintf('data%d', h); %// dynamic name
resultName = sprintf('res%d', h); %// dynamic name
base.(resName) = myFunction( base.(currentName) ); %// process data and save to result
end
The nice thing about this approach (especially if you run into memory problems) is that save and load supports this approach:
for h=1:n
dataName = sprintf('data%d', h); %// dynamic name
base = load( 'myHugeMatFile.mat', dataName ); %// loads only one variable from the file
%// now the variable is a field in base
resultName = sprintf('res%d', h); %// dynamic name
base.(resName) = myFunction( base.(currentName) ); %// process data and save to result
save( 'myResultsFile.mat', '-struct', '-append', 'base' ); %// please verify this works - I'm not 100% certain here.
end
Note how save and load can tread struct fields as different variables when needed.
base struct. For example, instead of data1 = [1,2,3]; use base.data1=[1,2,3]; and so on. now you can access using h: dn = sprintf('data%d',h); base.(dn)As far as I understand, you either want to make a matrix:
A = [1 2 3;
4 5 6;
7 8 9];
or a vector:
A =[1; 2; 3; 4; 5; 6; 7; 8; 9];
In the first case, just writing
A = [data1;data2;data3];
should do the trick. Otherwise, look into horzcat for a horizontal vector and vertcat for a vertical one:
A = horzcat(data1,data2,data3);
A = vertcat(data1',data2',data3');
A is not a matrix! And it's better to use .' when transposing (notice the dot). And it doesn't look remotely close to solving OP's problem, but hey... points for trying.; in the end of each row in the matrix is not necessary. This is to allow copy-paste of tables... (But they don't do any damage, so it's not a problem to have them there,,, it's just not something you have to do) =)
sprintfandfprintf. I could mentioneval, but then I would get comments sayingevalis evil!evalis evil!!! this is why you get these comments! use struct with dynamic field names instead