Thread Subject: textread - recall last position

Is it possible for textread to read from the last poistion it read
from? For file -

%test
10,20,30
40,50,60

I can read via:

[x1,x2,x3] = textread('t.txt','%d%d%d',
1,'delimiter',',','commentstyle','matlab');
[y1,y2,y3] = textread('t.txt','%d%d%d',
1,'delimiter',',','commentstyle','matlab', 'headerlines', 2);

Is there another solution which avoids using headerlines? Something
like fseek/ftell?

Thanks,
Sanjeev

On Feb 12, 8:22=A0am, supan...@yahoo.com wrote:
> Is it possible for textread to read from the last poistion it read
> from? For file -
>
> %test
> 10,20,30
> 40,50,60
>
> I can read via:
>
> [x1,x2,x3] =3D textread('t.txt','%d%d%d',
> 1,'delimiter',',','commentstyle','matlab');
> [y1,y2,y3] =3D textread('t.txt','%d%d%d',
> 1,'delimiter',',','commentstyle','matlab', 'headerlines', 2);
>
> Is there another solution which avoids using headerlines? Something
> like fseek/ftell?
>
> Thanks,
> Sanjeev

Use textscan for this

supandey@yahoo.com
> %test
> 10,20,30
> 40,50,60
> I can read via:
> [x1,x2,x3] = textread('t.txt','%d%d%d',
> 1,'delimiter',',','commentstyle','matlab');
> [y1,y2,y3] = textread('t.txt','%d%d%d',
> 1,'delimiter',',','commentstyle','matlab', 'headerlines', 2);
>
> Is there another solution which avoids using headerlines...

one of the solutions

     fnam='t.txt'; % <- your text file...
     s=textread(fnam,'%s','delimiter','\n');
     d=cellfun(@(x) sscanf(x,'%g,').',s(2:end),'uni',false);
     d{1:2} % <- or d{:}
%{
% ans =
          10 20 30
% ans =
          40 50 60
%}
% it comes with the benefit of NOT having too many vars in your ws...

us

On Feb 11, 1:25=A0pm, NZTideMan <mul...@gmail.com> wrote:
> On Feb 12, 8:22=A0am, supan...@yahoo.com wrote:
>
>
>
>
>
> > Is it possible for textread to read from the last poistion it read
> > from? For file -
>
> > %test
> > 10,20,30
> > 40,50,60
>
> > I can read via:
>
> > [x1,x2,x3] =3D textread('t.txt','%d%d%d',
> > 1,'delimiter',',','commentstyle','matlab');
> > [y1,y2,y3] =3D textread('t.txt','%d%d%d',
> > 1,'delimiter',',','commentstyle','matlab', 'headerlines', 2);
>
> > Is there another solution which avoids using headerlines? Something
> > like fseek/ftell?
>
> > Thanks,
> > Sanjeev
>
> Use textscan for this- Hide quoted text -
>
> - Show quoted text -

Unfortunately my version of Matlab (v6.5) does not have textscan.

supandey@yahoo.com wrote in message <b9f14884-002b-4cfe-b5ee-751ed6a1c3e3@w39g2000prb.googlegroups.com>...
[..]
>
> Unfortunately my version of Matlab (v6.5) does not have textscan.

You may give txt2mat from the file exchange a try - look at the doc for the 'FilePos' argument. It is compatible with v6.5.

supandey@yahoo.com
> Unfortunately my version of Matlab (v6.5) does not have textscan...

well, did you try other solutions that were presented to you...

us

On Feb 11, 4:23=A0pm, "us " <u...@neurol.unizh.ch> wrote:
> supan...@yahoo.com
>
> > Unfortunately my version of Matlab (v6.5) does not have textscan...
>
> well, did you try other solutions that were presented to you...
>
> us

I did not try this yet:

     fnam=3D't.txt'; % <- your text file...
     s=3Dtextread(fnam,'%s','delimiter','\n');
     d=3Dcellfun(@(x) sscanf(x,'%g,').',s(2:end),'uni',false);
     d{1:2} % <- or d{:}

The issue is that the file I want to read has about 40 million rows in
it (each row has 200 variables). I think the

s=3Dtextread(fnam,'%s','delimiter','\n');

will read the whole file in and than run into memory overflow
problems. Hence my need to read the file in chunks.

Thanks,
Sanjeev