Mixed Data Types In a Table

I'll preface this by stating I'm new to Matlab.

I have data stored in a table, and that table contains numeric (double), categorical, and ordinal data. I need to convert the table to a matrix of type "double" to run predictive classification (fitensemble). If I try the command: double(X), I get an error that "Conversion to double from table is not possible". I can't convert the table to an array, because its not possible to concatenate double and categorical arrays.

I've tried:

 classreg.regr.modelutils.predictormatrix (X);

which works, but drops the last column of the table.

How can I (efficiently) create a matrix of predictors from a table of mixed data types?

Thanks.

Answer 1

[Sean de Wolski]

% Simulate data
T = table(rand(10,1),categorical(cellstr(('rgbbrgrrbr').')))

% Convert to double and extract with {}
Td = varfun(@double,T)
mat = Td{:,:}