error problem with extracting data

Hi everyone.

I'm very glad to join you on this incredible community...

I'm trying to write a program to training and testing database for my machine learning project.

in first step I wrote a m-file to export files.

here is the code:

function erpsplit(indir)

d = dir(indir);
filelist = {};
datadirection = {};
classname = {'dislike','like'};
observations = {};

for i = 1:length(d)
    
   a = strcmp(d(i).name,classname{1});
   b = strcmp(d(i).name,classname{2});
   p = true;
   q = true;
   
   for m = 1:length(a)
       p = p & a(1,m);
   end
   
   for m = 1:length(b)
       q = q & b(1,m);
   end
   
   if (d(i).isdir == 0)
      continue;
   else if (p | q)
      datadirection = cat(2,datadirection,{[indir '\' d(i).name]});
   end
end

for k = [1,2]
    d = dir(datadirection{k});
    for i = 3:length(d)
       if (d(i).isdir == 0)
           filename = sprintf('%s\\%s',datadirection{k},d(i).name);
           filelist{i-2,k} = filename;
       end
    end
end

for k = [1,2]  
    for i = 1:16
        for j = 1:length(filelist)
            data = load(filelist{j,k});
            variables = fields(data);
            observations{j,i,k} = data.(variables{1});
            observations{j,i,k} = observations{j,i,k}(i,:);
        end
    end
end

s = 'Output Files';
mkdir(indir,s);
s = [indir '\' s];

for k = [1,2]
    mkdir(s,classname{k});
    s = [s '\' classname{k}];
    for i = 1:16
        str = ['ch#' num2str(i)];
        mkdir(s,str);
        s = [s '\' str];
        for j = 1:length(filelist)
            str = [classname{k} '_subject#' num2str(j) '_ch#' num2str(i) '.xls'];
            var = observation{j,i,k};
            s = [s '\\' str];
            save(s,'var');
        end
    end
end
        
    

end    

but an error occured during decoding the following line:

    d = dir(datadirection{k});

it says:

Index exceeds array bounds.

Error in erpsplit (line 32)

d = dir(datadirection{k});

but I tracing the code and the "datadirection" cell looks just fine.

Thanks everyone for assisting!

Answer 1

[Image Analyst]

At line 32, insert these lines:

whos k
whos datadirection

What do you see in the command window? Chances are datadirection has fewer cells than k.

[Hossein Shahbodaghkhan]

I checked the result after putting "datadirection" in an export variable for the function and it returned a 1*2 cell with my opinion values.but when I use your code it says that "datadirection" is an empty cell!

I don't know what's wrong with it...and of course I did these in line 32.

[Image Analyst]

Well, step through it line by line in the debugger and figure out why you never hit a line that assigns it to anything.

If you still can't figure it out, attach some files that your load() statement wants to read in and we'll do the same thing that you can do for you (that is, step through it line by line).

[Hossein Shahbodaghkhan]

Thank you very much for your assisting!

I finally found the error's cause and I made alot of changes for safty execution.

I attach my final code for everyone to use it if they need the same algorithm.

Answer 2

[Hossein Shahbodaghkhan]

the final code:

function erpsplit(indir)

warning('off');
d = dir(indir);
filelist = {};
datadirection = {};
classname = {'dislike','like'};
observations = {};

for i = 1:length(d)
    
   a = strcmp(d(i).name,classname{1});
   b = strcmp(d(i).name,classname{2});
   p = true;
   q = true;
   
   for m = 1:length(a)
       p = p & a(1,m);
   end
   
   for m = 1:length(b)
       q = q & b(1,m);
   end
   
   if (d(i).isdir == 0)
      continue;
   end
   if (p | q)
      datadirection = cat(2,datadirection,{[indir '\' d(i).name]});
   end
end

for k = 1:2
    d = dir(datadirection{k});
    for i = 3:length(d)
       if (d(i).isdir == 0)
           filename = sprintf('%s\\%s',datadirection{k},d(i).name);
           filelist{i-2,k} = filename;
       end
    end
end

for k = 1:2  
    for i = 1:16
        for j = 1:length(filelist)
            data = load(filelist{j,k});
            variables = fields(data);
            observations{j,i,k} = data.(variables{1});
            observations{j,i,k} = observations{j,i,k}(i,:);
        end
    end
end

s = 'Output Files';
mkdir(indir,s);
s = [indir '\' s]; dir1 = s;
cd(s);

k = 1;
while (k <= 2)
   a = strcmp(pwd,dir1);
   p = true;
   
   for m = 1:length(a)
      p = p & a(1,m);
   end
   
   if (p)
       mkdir(s,classname{k});
       s = [s '\' classname{k}]; dir2 = s;
       cd(s);
   else
       s = dir1;
       cd(s);
       continue;
   end
   i = 1;
   while (i <= 16)
       str = ['ch#' num2str(i)];
        a = strcmp(pwd,dir2);
        p = true;
  
       for m = 1:length(a)
           p = p & a(1,m);
       end
       if (p)
           mkdir(s,str);
           s = [s '\' str]; dir3 = s;
           cd(s);
       else
           s = dir2;
           cd(s);
           continue;
       end
       for j = 1:length(filelist)
           str = [classname{k} '_subject_n' num2str(j) '_channel_n' num2str(i) '.mat'];
           var = observations{j,i,k};
           f = [s '\\' str];
           save(f,'var');
       end
       i = i+1;
   end
   k = k+1;
end
        
   

end