MATLAB: Importing tsv files yyyy.mm.dd.tsv

arraydata importimportimporting excel dataloadmatricesmatrixmatrix arraymatrix manipulationxlswrite

Hi I'm beginner in MATLAB. I need to import tsv files that are named as follows:
2004.01.01.0000.tsv
2004.01.01.0300.tsv
2004.01.01.0600.tsv
2004.01.01.0900.tsv
2004.07.01.1200.tsv
2004.01.01.1500.tsv
2004.01.01.1800.tsv
2004.01.01.2100.tsv
2004.01.02.0000.tsv
2004.01.02.0300.tsv
2004.01.02.0600.tsv
2004.01.02.0900.tsv
2004.01.02.1200.tsv
2004.01.02.1500.tsv
2004.01.02.1800.tsv
2004.01.02.2100.tsv
...
...
2005.12.31.2100.tsv
so each day has 8 files associated with it.
I need to import them as a 3 dimensional matrix using for loops.
I have the idea that my code should look something like this but it is not easy as I think
for yy=2004:2005
y= int2str(yy);
  for mm=1:12
m= int2str(mm);
    for dd=1:31
d= int2str(dd);
      for ii=0000:0300:2100
i=int2str(ii);
z(:,:,)=importdata(y.m.d.i);
end
 end
  end
   end
Thank you.

Best Answer

yearvals = 2004:2005;
monvals = 1:12;
dayvals = 1:31;
hourvals = 0000:0300:2100;
for ynum = 1 : length(yearvals)
  y = yearvals(ynum);
  for mnum = 1 : length(monvals)
    m = monvals(mnum);
    for dnum = 1 : length(dvals)
      d = dvals(dnum);
      for hnum = 1 : length(hourvals)
        h = hourvals(hnum);
        filename = sprintf('%04d.%02d.%02d.%04d', y, m, d, h);
        if exist(filename, 'file')
          z(:, :, :, hnum, dnum, mnum, ynum) = importdata( filename );
        end
      end
    end
  end
end
... and better hope that the arrays are all the same size.
I ordered the indices as hour, day, month, year, because MATLAB is most efficient when the index that changes faster is to the left. You might find that for processing purposes, that a different arrangement is more efficient.
I should warn that with the above set-up, the non-existent days such as February 30th will have slots in the resulting array, but that zeros will be stored there. You can change that to another value such as NaN by putting an "else" on the "if exist". (That would fail, though, if it was initial files that were missing, such as if the data started from January 3rd because of trading holidays.)
Related Question