Create && Fill array variables resulting from the concatenation of string characters and numbers

Hi

I have been unsucessfully looking for an advice on a code that loops through a dataset (of cell type) and extracting each column as data vector, and hope MATLAB pundits could assist owing to the challenging feature of this task.

[i,j]=size(testdata);
k=2;
while k<=j % looping through columns
['_' num2str(k-1)'yr']=cell2mat(testdata(:,k)); %extract data from each column in "testdata "and assign it to variables "_kyr" as column vector
k=k+1;
end
--------------------------

Error: An array for multiple LHS assignment cannot contain LEX_TS_STRING.

Thanks in advance and regards

[Stephen23]

You are trying to create variable names dynamically. This is a bad idea, read the answers to know why.

Answer 1

[Stephen23]

https://www.mathworks.com/help/matlab/matlab_prog/string-evaluation.html

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

[Isma]

Thanks for the documentation which i hope won't create additional confusion than a concise feedback on the code. I simply wanted to read a term structure of interest rates (1 year, 2 years etc) from a big csv file before applying data reduction techniques. Otherwise I would have not bothered as "dynamically named variables" seems a recurent topic.Cheers anyway

[Stephen23]

Just use cell2mat on your whole cell array, you do not need to use separate variables. Then use indexing to quickly access the rows/columns that you require.

The links that I gave will not create "additional confusion", because they all state that creating variables names dynamically is slow, buggy and a bad idea. It is "recurrent topic" because beginners keep thinking that it is a great idea.

[Isma]

Thanks. This was my idea of last resort: apply cell2mat on whole array before carrying on, but will keep going through the links to see how bad 'eval' really is to the extent that it appeared to be the solution to my data cleansing. Cheers

Thanks for the links