can someone help me with this svmtrain's error "Error using svmtrain ,TRAINING must be a numeric matrix"?

I am trying to figure out why i am seeing this error but i am stuck, here is the code:

[X,T] = catadataset;
 xdata = X;
 group = T;

   svmStruct1 = svmtrain(xdata,group,'kernel_function');
   species = svmclassify(svmStruct1,feat,'showplot',false)

catadataset:

function [X,T] = catadataset
  datastruct_x = load('csvlistall.mat');
  X = datastruct_x;
  datastruct_t = load('cata_Label1.mat');
  T = datastruct_t;

end

csvlistall.mat is a 88X95 matrix, and it contains only numbers ,cata_Label1.mat is a 1X95 matrix and also contains only numbers. can somebody help me please?

[Walter Roberson]

Please show class(X), class(T)

Because you are loading from a .mat file, there is the possibility that the .mat file is a true binary .mat file instead of being a text file. A binary .mat file would return a struct() from load() rather than a numeric matrix, even if the .mat file only contains one variable.

It might end up looking something like

function [X,T] = catadataset
  datastruct_x = load('csvlistall.mat');
  X = datastruct_x.csvlistall;
  datastruct_t = load('cata_Label1.mat');
  T = datastruct_t.cata_Label1;
end

[Hadeer tawfik]

yes this is true, it was a true binary .matfile, so I have changed the code into yours and it has worked perfectly , thank you so much