Thread Subject: load txt with different row length

Hallo,
I have an inputfile which is stored as a txt tab-delimited.
The problem is that the input consists of n columns and m
rows where each row has not the same number of columns. How
can I load this into matlab if possible in a cell array.
How can I than sort each row?

Here is an example:

Input:
1 o1.2 ABC o6.5 GAD o1.6
2 GAD o1.9 o6.3

sorted output:
1 o1.2 o1.6 o6.5 ABC GAD
2 o1.9 o6.3 GAD

How can I than store this as a txt file?


Thanks, Corinna

> Input:
> 1 o1.2 ABC o6.5 GAD o1.6
> 2 GAD o1.9 o6.3
>
> sorted output:
> 1 o1.2 o1.6 o6.5 ABC GAD
> 2 o1.9 o6.3 GAD
>
> How can I than store this as a txt file?
>
>
> Thanks, Corinna

You should have a look at the loadcell tool on the matlab file exchange
server:

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=1965&objectType=file

This is optimized for tables though, so I guess that you need to
transform the resulting 2D cell array to a 1D cell array afterwards (and
remove the empty or zeros fields). An alternative is the readtext function:

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=10946&objectType=FILE

Kris

Hallo Kris,

sorry, I cannot run those programs. Once the pc says he
cannot find the matlab-file but I copied it in the right
folder. otherwise he said input is wrong. Can you give me
another solution.

The input is a txt-delimited files and the delimiter
between entries is a tab.

> > Input:
> > 1 o1.2 ABC o6.5 GAD o1.6
> > 2 GAD o1.9 o6.3
> >
> > sorted output:
> > 1 o1.2 o1.6 o6.5 ABC GAD
> > 2 o1.9 o6.3 GAD
> >
> > How can I than store this as a txt file?

Any idea?

Corinna

> otherwise he said input is wrong.

What do you mean? Which function fails? I always use something like:
[data,b,numdata]=loadcell ('c:\data.txt', ['"' char(9)], '');

or
[data,b,numdata]=loadcell('c:\data.txt', ['"' char(9)], 'single-string')

%convert input to strings
[p1,p2]=find(isnan(numdata) == false);
for i=1:length(p1)
data{p1(i),p2(i)} = num2str (data{p1(i),p2(i)});
end

%sorted output (2D cell array)
for i=1:size(data,1)
     p=find (numdata(i,:) == 0); %find the empty items
     tmp=data(i,:);
     tmp(p)=[];
     data2{i}=sort(tmp);
end


> The input is a txt-delimited files and the delimiter
> between entries is a tab.
>
>>> Input:
>>> 1 o1.2 ABC o6.5 GAD o1.6
>>> 2 GAD o1.9 o6.3
>>>
>>> sorted output:
>>> 1 o1.2 o1.6 o6.5 ABC GAD
>>> 2 o1.9 o6.3 GAD
>>>
>>> How can I than store this as a txt file?

[f,msg] = fopen('c:\data2.txt', 'wb', 'l');
for i=1:size(data2,2);
fprintf(f, '%s\t', data2{i}{:})
fprintf(f, '\n');
end
fclose(f);

Hallo Kris,

I found another solution. Now I tried to store my
resultfile (dimension 8x11 cell) as a tab-delimited
textfile with save but it doesn't work.

EDU>> save 'ergebnis.txt' 'result' -ascii
Warning: Attempt to write an unsupported data type to an
ASCII file.
Variable 'result' not written to file.

Also your solution does not work
  
> [f,msg] = fopen('c:\data2.txt', 'wb', 'l');
> for i=1:size(data2,2)
> fprintf(f, '\n');
> end
> fclose(f);

Have you got an alternative?

Thanks, Corinna

Hallo,

here is the error message:

EDU>> [f,msg] = fopen('data2.txt', 'wb', 'l');
EDU>> for i=1:size(result,2)
fprintf(f, '%s\t', result{i}{:})
fprintf(f, '\n')
end
??? Cell contents reference from a non-cell array object.


What means this?

Corinna

In article <fmqia9$heh$1@fred.mathworks.com>,
Corinna Schmitt <csc@igb.fhg.de> wrote:

>here is the error message:

>EDU>> [f,msg] = fopen('data2.txt', 'wb', 'l');
>EDU>> for i=1:size(result,2)
>fprintf(f, '%s\t', result{i}{:})
>fprintf(f, '\n')
>end
>??? Cell contents reference from a non-cell array object.

>What means this?

You have not given a value to result. You open the file but
you do not read from the file.
--
We regret to announce that sub-millibarn resolution bio-hyperdimensional
plasmatic space polyimaging has been delayed until the release
of Windows Vista SP2.

On Jan 18, 10:50=A0am, rober...@ibd.nrc-cnrc.gc.ca (Walter Roberson)
wrote:
> In article <fmqia9$he...@fred.mathworks.com>,
>
> Corinna Schmitt <c...@igb.fhg.de> wrote:
> >here is the error message:
> >EDU>> [f,msg] =3D fopen('data2.txt', 'wb', 'l');
> >EDU>> for i=3D1:size(result,2)
> >fprintf(f, '%s\t', result{i}{:})
> >fprintf(f, '\n')
> >end
> >??? Cell contents reference from a non-cell array object.
> >What means this?
>
> You have not given a value to result. You open the file but
> you do not read from the file.
> --
> We regret to announce that sub-millibarn resolution bio-hyperdimensional
> plasmatic space polyimaging has been delayed until the release
> of Windows Vista SP2.


Maybe something like this? This one reads all odd rows with 5
arguments and even rows with even rows with 8. If there is a pattern
to how many arguments are in each row, this should work fine.

fid =3D fopen('filename.txt');
for i =3D 1:2:362
    rawdat{i,:} =3D textscan(fid,'%n %n %n %s %s',1);
    rawdat{i+1,:} =3D textscan(fid,'%n %n %n %n %n %n %n %s',1);
end
fclose(fid);