wildcard in filename for readtable does not work

When trying to load the table in matlabwith table = readtable("...\SP17\2022SP17.xlsx"), I can get the file loaded, but when I try to do it with a wildcard: table = readtable("...\SP17\2022*.xlsx") (or another variation with an asterix) it gives the error:

Error using readtable

Unable to find or open '...\SP17\2022*.xlsx'. Check the path and filename or file permissions.

I need to use the wildcard later as I don't know the full name of the files, I need to use.

[Stephen23]

"when I try to do it with a wildcard: table = readtable("...\SP17\2022*.xlsx") "

Can you point to the exact location in the READTABLE documentation where it states that the filename can include wildcard characters?

"I need to use the wildcard later as I don't know the full name of the files"

Use DIR:

https://www.mathworks.com/help/matlab/import_export/process-a-sequence-of-files.html

Answer 1

[Alexander]

If I understood your question correctly, I would do it this way:

Files = ls('.\SP17\2022*.xlsx')
nFiles = size(Files); nFiles = nFiles(1);
for ii = 1:nFiles
    tmpFile = strtrim(Files(ii,:))
    table = readtable(tmpFile)
    % Do your computation
end

I haven't tested this short code, but anyway the principle should be clear.

Edited after comment from Stephen23.

[Stephen23]

This code is buggy for due to the rather poor choice of LS rather than the usual DIR, e.g.:

• LS returns a character array where each row contains one or more file/folder names. So Files(ii) will be the first character of each row, which is very unlikely to be useful or correct.
• In general file/folder names can have different lengths, but the output from LS is left-aligned and right-padded with spaces to the length of the longest file/folder name. So to handle the names robustly requires not only splitting them from the character array but also stripping those extra trailing space characters. At this point astute readers will notice that space characters are perfectly valid characters within file/folder names (and extensions are optional), so there is literally nothing that can distinguish between trailing space characters added by LS and trailing space characters within the name itself. Stripping them can distort the names so they do not match the actual names. Using LS like this is fundamentally fragile/buggy. There is no way to resolve this... except by easily avoiding the situation and writing better code using DIR.

Using DIR is also what the MATLAB documentation recommends:

https://www.mathworks.com/help/matlab/import_export/process-a-sequence-of-files.html

[Alexander]

It was a pull ahead answer, the code was not tested and as written above it should only show the principle. . I've edited the code should now run w/o error.

[Walter Roberson]

Files = ls('../..')

Files = 
    ' bd_build   boot	  Data	 etc    lib     MATLAB		 'MATLAB Drive'   mnt   proc	 root   sbin   SupportPackages	 tmp	 usr
      bin	    CourseFiles   dev	 home   lib64  'MATLAB Add-Ons'   media		  opt   public	 run    srv    sys		 users	 var
     '

You can see that on MacOS and Linux, ls returns a character array with many different files on the same line -- unless, that is, you use the -1 option

Files = ls('-1', '../..')

Files = 
    'bd_build
     bin
     boot
     CourseFiles
     Data
     dev
     etc
     home
     lib
     lib64
     MATLAB
     'MATLAB Add-Ons'
     'MATLAB Drive'
     media
     mnt
     opt
     proc
     public
     root
     run
     sbin
     srv
     SupportPackages
     sys
     tmp
     users
     usr
     var
     '

Answer 2

[Les Beckham]

You must call readtable with a valid filename for an existing file. It will not accept wildcards.

You could use uigetfile to open a dialog that will allow you to pick the file. It will return the file path and name. Then use fullfile to create the full filename, including path, that you can pass to readtable.

[Cris LaPierre]

The problem with a wildcard is that there is no guarantee it returns a single file. You could use wildcards to collect file names using the dir function or datastores, and then use the result of that to load files.