Automated concatenation of variables into matrix then separation of matrix into variables containing original variable names

Hi,

I have a .mat file with several variables listed in the Workspace. I would like to 1) concatenate the variables into a matrix where each variable is a column, 2) convert the columns back to variables with their original names. Since I don't always know the variable names or the number of variables in the workspace, I need this process to be automated. I was thinking of doing something like this:

origvarnames = who;

for i = 1:length(origvarnames)

data(:,i) = ?variable?(i)

end

% where ?variable? is the nth variable in the workspace.

% after some work is done to the values within 'data' the columns need to go back into variables

for j = 1:length(origvarnames)

?variablename? = data(:,j)

end

% where the variablename could maybe come from the 'origvarnames' variable?

Could anyone fill in blanks or inform me of any useful functions for this process? If I am way off with my method then please tell me how you would carry out this procedure.

Thanks,

Dylan

[Matt Fig]

What is the point of converting them to a matrix, then converting them back? What do you expect to gain from this at the end? In other words, if you already have them as separate variables, why go through the two steps to get them back as separate variables??

Answer 1

[Fangjun Jiang]

It can be done but I am not sure if this is really necessary to solve your original problem. The following example assumes all your data is column vector (nx1 array). Step through to see the effect and see if you understand every one of them.

clear all;
a=(1:10)';
b=(100:10:150)';
c=rand(3,1);
Vars=whos;
data=[];
for k=1:length(Vars)
    data=[data;eval(Vars(k).name)];
end
data=2*data;
for k=1:length(Vars)
    assignin('base',Vars(k).name,data(1:Vars(k).size(1)));
    data(1:Vars(k).size(1))=[];
end

[Dylan Finley]

This worked great, just had to make some minor adjustments to the script to fit it into my process and ensure that it assigned the data in a column format instead of a row format. Thank you for your help!

Answer 2

[Paulo Silva]

%without any verification of the format of all variables
a=[1 2 3]
b=[4 5 6]
lvar=who;
c=cell2mat(cellfun(@eval,lvar,'uni',false))'

Answer 3

[Walter Roberson]

Are all of the matrices the same length? Could cell matrices be used?

The most natural approach would seem to be to use a structure with dynamic field names: is there reason not to use that?