How to glean the files I want from so many data files?

Dear MATLAB Experts,

I have had this problem for a couple of days and I could not think of a way to resolve this. I do appreciate any helps to get my way out!

I have about 100000 files which names are in the following format:

 sprintf ("%dfile%d.csv" , identifier 1, identifier 2)

where identifier 1 is numbers 1, 2, 3, 4, 5 ..., 50 (in order) and identifier 2 could be any number. Some examples are

 1file1234.csv
 1file2003.csv
 1file11111111.csv
 2file6667.csv
 2file99999.csv
 3file1.csv
 3file10000.csv
 .
 .
 .
 50file3456.csv
 50file123456.csv

The files I am interested in are the ones meet this condition:

for each identifier 1 from (1 to 50), the file with smallest identifier 2 is of interest. All other should be deleted. In the above examples, we need

 1file1234.csv
 2file6667.csv
 3file1.csv
 .
 .
 .
 50file3456.csv

How can I write a code in MATLAB does this for me?

Thank you so much in advance.

Answer 1

[Image Analyst]

Use dir() on the folder to get the filenames. Then use sscanf() to extract the two numbers. Easy, but let us know if you can't do it.

[Homayoon]

unfortunately could not. I used

a = dir *file*.csv
sscanf (a(1,1) , '%d' , '%d')

but did not work out.

[Walter Roberson]

For example,

>> S = '2file6667.csv'; sscanf(S, '%dfile%d.csv')
ans =
           2
        6667

[Homayoon]

Thanks Walter.

What you have suggested is really appreciated but does not do what I seek. Once I am using dir as in

filelist = dir(fullfile( '*file*.csv'));

it gives me array of 10000*1 where 10000 is the number of files in my directory. and then I did

sscanf(filelist(1,1), '%dfile%d.csv')

it does not work out.

Beside this, I need the procedure to be automatically applied to all files in the directory not just a single file. I am just unable of finding my way out.

[Walter Roberson]

dinfo = dir(fullfile( '*file*.csv'));
filelist = {dinfo.name};

bestids = inf(1,50);
bestnames = cell(1,50);
for K = 1 : length(filelist)
  this_file = filelist{K};
  ids = sscanf(this_file, '%dfile%d.csv');
  oldbest = bestids(ids(1));
  if ids(2) >= oldbest
    delete(this_file);
  else
    oldbestname = bestnames{ids(1)};
    if ~isempty(oldbestname)
      delete(oldbestname);
    end
    bestids(ids(1)) = ids(2);
    bestnames{ids(1)} = this_file;
  end
end

At each point, if the id2 is not the best we have seen for this id1 then delete the current file; if it is better than our previous best then delete the previous best and keep the one we just encountered.

[Homayoon]

Thank you so much Walter. I now understand the code. Thanks.