pick files within dir

I have a list create by dir

i want to pick specific files in the dir the name differentiators are genreal name with "post" on the title and without

I want to pick "post" files to create new table with the

files = dir(test{2});
match = readnames(files.name,'Post'); %I know this is wrong but im not sure hwat else to try
comp.files = match;
x =  size(comp.files);

help please

[Ihaveaquest]

it all looks great thak you i figured it out

files = dir(fullfile('*Comp*Post*'));

this si what i was looking for .. thank sorry for the misunderstanding

Answer 1

[Walter Roberson]

dinfo = dir('*post*');
comp.files = fullfile({dinfo.folder}, {dinfo.name});
numfiles = length(comp.files);

Now comp.files will be a cell array of character vectors, each of which is a fully-qualified file name that has "post" as part of the name.

For other files without "post" the code is slightly longer

npdinfo = dir('*');
ndpinfo([ndpinfo.isdir]) = [];  %remove . and .. and any folder names
npdinfo(contains({npdinfo.name}, 'post')) = [];   %remove names containing 'post'
comp.npfiles = fullfile({npdinfo.folder}, {npdinfo.name});
numnpfiles = length(comp.npfiles);

... If you are wanting both lists, then it is easier to just do one pass, like

dinfo = dir('*');
dinfo([dinfo.isdir]) = [];   %remove . and .. and any folder names
mask = contains({dinfo.name}, 'post');
comp.files = fullfile({dinfo(mask).folder}, {dinfo(mask).name});
numfiles = length(comp.files);
comp.npfiles = fullfile({dinfo(~mask).folder}, {dinfo(~mask).name});

[Ihaveaquest]

so due to my complete code using in other areas

files = dir(test{2});

i need to add the add ons here, after ^^^

match being the function to modify to pick between the post files

match = readnames(files.name,'Post'); %I know this is wrong but im not sure hwat else to try

comp.files = match;

x = size(comp.files);

would your code do that for me ??

[Ihaveaquest]

it worked but it picked other post files that i did not realized where there and did not want. how may i pick just the comp post files ?

[Walter Roberson]

I do not know what you mean by "comp post files" ?

I do not know what "readnames()" is.

I will assume that test{2} is the name of a folder.

I am not clear whether you are looking for 'Post' or 'post' in the name ?

dinfo = dir(fullfile(test{2}, '*'));
dinfo([dinfo.isdir]) = [];   %remove . and .. and any folder names
mask = contains({dinfo.name}, 'post');
comp.files = fullfile({dinfo(mask).folder}, {dinfo(mask).name});
numfiles = length(comp.files);
comp.npfiles = fullfile({dinfo(~mask).folder}, {dinfo(~mask).name});

[Ihaveaquest]

would you explain this line - what it does

ndpinfo([ndpinfo.isdir]) = []; %remove . and .. and any folder names

[Ihaveaquest]

i guess what i am saying is that theres two conditions that the dir has to read through "comp" and "post"

the files names are "alwentrtginrig_comp_kaejtnkjbawekjfgber_post"

[Walter Roberson]

The output from dir() is a struct. For your purposes, the important fields are 'isdir' (which is true if the directory entry is for a folder); and 'folder' (name of the containing folder); and 'name' (name of the file within the folder).

For any nonscalar structure, STRUCTURENAME.FIELDNAME undergoes "struct expansion". The result is the same as if you had coded

STRUCTURENAME(1).FIELDNAME, STRUCTURENAME(2).FIELDNAME, STRUCTURENAME(3).FIELDNAME, etc

-- as if you had coded all of that, commas included, as separate parameters to whatever function you are using. So if you use the [] function

[STRUCTURENAME.FIELDNAME]

the result is as-if you had typed

[STRUCTURENAME(1).FIELDNAME, STRUCTURENAME(2).FIELDNAME, STRUCTURENAME(3).FIELDNAME, etc]

which would create a list containing all of the FIELDNAME files for each entry in the non-scalar structure STRUCTURENAME

So [dinfo.isdir] creates a vector of values, the first of which is dinfo(1).isdir, the second is dinfo(2).isdir, and so on -- so [dinfo.isdir] is a vector that tells you whether each of the corresponding dinfo entries refers to a folder.

As this would be a logical vector, it can be used to select elements of an array. dinfo([dinfo.isdir]) is therefore subset of the dinfo struct such that the isdir entry was true for that entry.

Then when you assign [] to those entries, that is MATLAB syntax to delete the entries.

So after

ndpinfo([ndpinfo.isdir]) = [];

then ndpinfo will not contain any entries that deal with folders.

The further information that is relevant is that unless you are using a quite old filesystem type, then when you ask for a list of all entries in a folder, using '*' as the pattern, then you will not just be given the list of files. Instead, you will be given the list of files (and subfolders) and one entry named . and one entry named .. . The entry . refers to the directory itself, and the entry .. refers to the parent directory. These turn out to be quite handy to use for various purposes, but they are not the names of files you happen to be looking for in your particular purpose, so you want to remove those from the list.

It is common but mistaken to believe that the . and .. entries will always be the first two entries returned by dir(), so it is not unusual for people to take dir() results and loop starting with entry #3. That is incorrect code in every operating system that MATLAB has ever been supported on. If you have a reason to avoid all directory entries including . and .. then use the isdir filtering that I show; if you just want to avoid . and .. specifically then filter for those based on name, not on assumptions about their relative positions.

[Walter Roberson]

dinfo = dir(fullfile(test{2}, '*'));
dinfo([dinfo.isdir]) = [];   %remove . and .. and any folder names
mask = contains({dinfo.name}, 'post') & contains({dinfo.name}, 'comp');
comp.files = fullfile({dinfo(mask).folder}, {dinfo(mask).name});
numfiles = length(comp.files);
comp.npfiles = fullfile({dinfo(~mask).folder}, {dinfo(~mask).name});