fid=fopen(...
while ~feof(fid)
v=cell2mat(textscan(fid,'%f',9,'collectoutput',1));
% do whatever with _v_ here
...
end
fid=fclose(fid);
Alternatively, unless the files are huge, it'll be faster to read the array and process it from memory rather than reading the file record-by-record. (Plus, you may well be able to vectorize the "whatever" and avoid any looping altogether, almost always agoodthing (tm) in Matlab to strive for...
In that case,
fid=fopen(...
a=cell2mat(textscan(fid,'','collectoutput',1));
fid=fclose(fid);
% process the array here...
NB: Two things--first, note that wrapped the textscan call inside cell2mat to return a double array instead of a cell array; when the data are all numeric is much simpler addressing and less overhead. Second, in the second case the use of the empty format string is a flag to textscan that it will return the array in the shape found in the file automagically--you don't need to know either number of records nor items per record as long as it is a regular array.
