how to put vectors with different size in one matrix and how to suppress the NAN in this matrix

Hello, I have two difficulties with my script, the first is that I have vectors with different size et I want to create a matrix and then to fill with my different vectors but I think I do the wrong thing with Mat=[ 10000,11] . The other thing is that I want clearly to suppress all NAN in my matrix only in my vector TR_NAN. Have you an idea please?

filename =  'C:\Users\williot\Dropbox\UQTR 2013 - 2015\E-Prime\exp 2 ÉNPQ\eye-tracking\données brutes 30 aspirants\test_matlab.xlsx';
%'E:\exp 2 ÉNPQ bis\Cueing + TRV\test_script_matlab.xlsx';
[num,txt,raw] = xlsread(filename);
%--création matrice Mat---------------------------------------------
Sujets = num(:,1);
Essais = num(:,30);%82
TR_NAN = num(:,115);%104
TR_acc = num(:,112);%101
validite = txt(:,147);%138
contenu = txt(:,112);%101
question = txt(:,143);%134
procedure = txt(:,137);%68
bloc = txt(:,95);%84
running = txt(:,92);%137
Data_OK = [];
% Mat = [1009 138]
Mat=[]
Mat=[10000,11]
Mat = [Sujets Essais TR_NAN TR_acc Data_OK validite  ...
       contenu question procedure bloc running];
Mat(isnan(Mat(:,3)),:) = [];
%===================================================================
%--exclusion des TR à + ou - 2ET------------------------------------
for suj = 1 :length (Sujets);
  Moy_TR = mean (TR_NAN);
  ET_TR = std(TR_NAN);
  Data_OK = TR_NAN((TR_NAN>Moy_TR-2*ET_TR) & (TR_NAN>Moy_TR+2*ET_TR));
end
%--CONSTRUCTION D'UN FICHIER DE SORTIE-------------------------
%Création de la Matrice finale------------------------------------------------
Mat_out = [Sujets Essais TR_NAN TR_acc Data_OK validite ...
           contenu question procedure bloc running];
%--nom fichier------avec excel titre col---------------------
path_and_name_file='PREP_extremes.txt';
%--ouverture fichier
fid=fopen (path_and_name_file,'wt');
%titre colonne-----------------------------------------------
fprintf(fid,'Sujets;Essais;TR_NAN;TR_acc;Data_OK;validite;contenu; ...
                                    question;procedure;bloc;running');
fprintf(fid,'\n');
%écriture
for n=1:length(Mat_out);
  fprintf(fid,'%d;%d;%d;%d;%d;%s;%s;%s;%s;%s;%s',Mat_out(n,:));
     %1.4f;%1.4f;%1.4f;%1.4f;%1.4f;%1.4f;%1.4f;%1.4f;',Mat_out(n,:));
  fprintf(fid,'\n');
end
fclose(fid);

Answer 1

[Alexandre Williot]

Thank you, here what I needed,

Essais =[];
    Essais = num(:,93);
    Essais(isnan(Essais))=0;
    Pos_OK=find(Essais ~= 0);
    Pos_OK1=find(Essais == 0);
    num=num(Pos_OK,:);
    txt= txt(Pos_OK,:);
    raw= raw(Pos_OK,:);  

      Essais = num(:,93);

Answer 2

[dpb]

Matlab can only treat "ragged" arrays as a cell array. Use the "curlies" ( {} ) to place an array in a cell array location.

for ix=1:N
  C(ix)={a_vector};
end

C will be a cell array of length N arranged as a row vector by default.

I don't follow the question regarding the NaN, sorry.

[Walter Roberson]

TR_NAN = TR_NAN(~isnan(TR_NAN));