Creating a matrix of random numbers which their distributions are stored in a table?

i am trying to create a matrix of random numbers with each row following a certain probability distribution function (PDF) . also, PDFs are stored in another table such as this:

Code DistFit

____ _______________________

1057 'gamrnd(1.9948,5.231)'

1072 'wblrnd(1.1703,5.1103)'

1101 'wblrnd(1.0686,1547.6)'

1112 'wblrnd(2.0439,11.722)'

1114 'wblrnd(1.2258,14.977)'

for example i want to created a 5 * 30 matrix which its first row ( regarding the code "1057") follow a gamrnd(1.9948,5.231)

thank you beforehand

[Adam Danz]

Are you building this table or is it something that's provided and outside of your control? If you're building it, a much better organization would be to separatesly list the distribution type and inputs, in separate columns. Otherwise we need to first decompose the DistFit column into the function and the inputs which could potentially be sloppy.

[Masoud Askari]

Thank you for your answer Adam

the actual data is saved in an excel file and has 150 rows which i inputted into MATLAB using 'readtable'.

[Adam Danz]

In that case, my answer below should do the trick. Let me know if you have any questions.

Answer 1

[Adam Danz]

It would be better to construct the DistFit table so that the function and the input values are in separate columns. That way you don't have to deconstruct the DistFit value into the function and input parts. The answer below assumes you're working with the table in your answer but if you are producing that table and are able to create it such that the function and inputs are in separates columns, a much cleaner solution is available.

% Re-create the first 3 rows of the table in the question
Code = [1057 1071 1101]'; 
DistFit = { 'gamrnd(1.9948,5.231)' ,'wblrnd(1.1703,5.1103)', 'wblrnd(1.0686,1547.6)'}.'; 
T = table(Code, DistFit); 

% Define the size of the array and the code you want to look up
sz = [5,30]; 
code = 1101; 

% decompose distFit into function and inputs
dist = regexprep(T.DistFit(T.Code==code),'\(.+\)',''); 
inputs = regexprep(T.DistFit(T.Code==code),{'.+(',')'},{'',''}); 
inputVals = cellfun(@str2double,strsplit(inputs{:},','),'UniformOutput',false);
fcn = str2func(dist{1}); 

% Compute the distribution with the given size
data = fcn(inputVals{:},sz);