Re: Loading ASCII data files quickly

[Michael Grossbach]

Matthias Brennwald wrote:
> On Aug 10, 2009, at 5:03 PM, Christoph Ellenberger wrote:
>
> > I usually use sed to straighten the data and then do the load  
> > thing....
> > something like:
> >
> > cmd=['sed -n -e "/^ *[0-9]/p" ' fname  ">" Tfname];
> > if (system(cmd,1))
> >   error("Fehler 1");
> > end
> >
> > xx=load(Tfname);
> >
> >
> > where fname is the filename and tfname is some temporary filename  
> > which I remove afterword with....
> >
> > cmd=["rm " Tfname];
> > if (system(cmd,1))
> >    error("Fehler 4");
> > end
> >
> > but this depends on what structure do you want afterward. I  
> > sometimes have the coulumnames in the file as well and create a  
> > cellarray with the datapoints called Data.X, Data.Y etc. as well.
> >
> > Hope this helps
> > Christoph
>
> Ok, but this requires sed to be available. My code needs to work with  
> non-unixy machines, too.
Try reading the header lines using fgetl and the rest of the data with 
fscanf.
Not tested:
function tdm = tdmread(path2tdmfile)
[fid, msg] = fopen(path2tdmfile, 'r');
if fid < 3, error(msg), endif
first = fgetl(fid);
second = fgetl(fid);
data = fscanf(fid, '%f%f', Inf)
tdm = struct('header', {first; second}, 'data', data);
fclose(fid);

HTH, Michael