Dir command addresses one extra file

Hello all,

I have a folder containing 9 .htk files. I need to use "dir", and then "readhtk" in a loop to import them to MATLAB, but DIR appears to give 10 files instead of 9! here is my code:

    htkfiles = dir('/Users/Desktop/Acsegment/mfcdir/*.htk');
    nhtkfiles = length(htkfiles); % 10!!! It should be 9 tough!
    data = cell(nhtkfiles,2);
    for k = 1:nhtkfiles
        b(k,1) = strcat({'/Users/Desktop/Acsegment/mfcdir/'},{htkfiles(k,1).name});
        eval(['data{k,1} = readhtk(b{k,1});']);
    end

When looking at the filenames in htkfiles, I have them like this:

      htkfiles(1,1).name = '.htk'
      htkfiles(2,1).name = 'fadg0_si1279.htk'
      htkfiles(3,1).name = 'fadg0_si1909.htk'
      htkfiles(4,1).name = 'fadg0_si649.htk'
      htkfiles(5,1).name = 'fadg0_sx109.htk'
      htkfiles(6,1).name = 'fadg0_sx19.htk'
      htkfiles(7,1).name = 'fadg0_sx199.htk'
      htkfiles(8,1).name = 'fadg0_sx289.htk'
      htkfiles(9,1).name = 'fadg0_sx379.htk'
      htkfiles(10,1).name = 'faks0_si943.htk'

Comparing to what I see in that folder, the first file is not supposed to be there! Anyone got any ideas why Im getting one extra file?

[per isakson]

I cannot reproduce the issue on R2013a, Win7, 64bit. However, .htk seems to be a legal file name. Are you sure there isn't a file with that name on the disk.

[Guillaume]

As others have said, if dir returns that file that's most likely because the file is there.

However, I wamted to comment on your example. You don't need a loop to generate b:

b = strcat(''/Users/Desktop/Acsegment/mfcdir/', {htkfiles.name});

And more importantly, don't use eval to just call a function. You can also avoid the loop using cellfun, in which case you don't even need to predeclare data:

data = cellfun(@readhtk, b, 'UniformOutput', false);

Answer 1

[Guillaume]

If all you want to do is remove files starting with '.', it's simply

htkfiles = dir('/Users/Desktop/Acsegment/mfcdir/*.htk');
htkfiles(strncmp({htkfiles.name}, '.', 1)) = []; %remove files and dir starting with '.'

Or, to remove directory as well (a good idea):

htkfiles(strncmp({htkfiles.name}, '.', 1) | [htkfiles.isdir]) = []; %remove directory from list as well

As per my comment, you can then continue with:

htkfullpath = strcat('/Users/Desktop/Acsegment/mfcdir/', {htkfiles.name});
data = cellfun(@readghtk, htkfullpath, 'UniformOutput', false);

[Negar]

Thank you so much! It was a great idea! :)

Answer 2

[Orion]

Hello,

isn't there a hidden file in your folder ? ( '.htk' )

it's still weird because a file without a name and just an extension shouldn't be accepted by windows.

Otherwise, you can identify your files ?

htkfiles = dir('/Users/Desktop/Acsegment/mfcdir/fa*.htk');  % add _fa_ to find all .htk files beginning with it

[Negar]

Hi, Im not sure about hidden files, Im not working on windows, Im on Mac OS (UNIX), could it be therefor? regarding your last suggestion, I actually have many many files (1344), with various names, not all starting with 'fa'. its just an excerpt I have included here. But how could I see if there is some hidden files in that folder? and how to exclude them? Thank you!

[Negar]

But how can I filter out that filename which starts with '.'?

[Titus Edelhofer]

Hi,

files starting with . are filtered out if you do a "ls" in a shell. Try "ls -a" and you will see the file (or folder) .hks. If it's a folder, take a look inside, if it contains some configurations.

I suspect it's a file that was somehow accidentially created (probably empty). In this case delete it ;-).

Titus

[per isakson]

• On Windows the version control program, Mercurial, uses folders named .hg
• IIRC FEX-submissions from the mac-world includes folders, the names of which have a period in the first position. Possibly those folders are hidden on the mac. Yes, I'm guessing. Check whether htkfiles.isdir is true

[Negar]

 "ls -a" gives me 9 files, the .htk file is not included.