I have a simple script in matlab and I want to load a file. It seems it only works if the file is in the same dir as the script. If I add the file to a directory it does not read it.
For example:
fileID = fopen('myfile','r' ,'n', 'US-ASCII');
but when I put myfile in files:
fileID = fopen('files/myfile','r' ,'n', 'US-ASCII');
or
fileID = fopen('./files/myfile','r' ,'n', 'US-ASCII');
I get a -1 as a fileID. File cannot be read.
As per the comments, this is happening because you most likely added the path of where the script was located to your MATLAB path but you did not add the subdirectory where the file was in to your path. This is why it can't find the file. Therefore to avoid this in the future, you need to physically change the directory (i.e. the Working Directory) of where MATLAB is currently operating to where your script is stored.
It is then where local referencing should work. You can do this by either using the cd function, going to the top of your MATLAB window where you see the directory listing, clicking on the arrow to the right and pulling a drop down menu to change the directory, typing the actual directory you want by clicking on any blank space in the directory listing to enable a text box:
... or if you are running the code in the MATLAB editor, it'll request that you change directories as the script you are trying to run is not currently located in the working directory.
You can also programmatically add the subfolders in your script directory using mfilename, fileparts, genpath and addpath:
[dir, ~, ~] = fileparts(mfilename('fullpath')); % locate your script directory
addpath(genpath(fullfile(dir))); % add the folder and all subfolders to Matlab search directory
% then load your file.
fileID = fopen('myfile','r' ,'n', 'US-ASCII')
If it is also important that all outputs should be placed within the same directory as your script file, you can cd to your script directory:
cd(dir)
Related
I've just created a script file in MATLAB, but can not run it. The name of my script is getEnvFiles.m. When I first tried to run it, I got the following result:
>> getEnvFiles
'getEnvFiles' is not found in the current folder or on the MATLAB path, but exists in:
\\wsl$\ubuntu\home\me
Change the MATLAB current folder or add its folder to the MATLAB path.
So, I added this directory (which is actually the current directory) to the search path, but still got the same result:
>> addpath('\\wsl$\ubuntu\home\me')
>> getEnvFiles
'getEnvFiles' is not found in the current folder or on the MATLAB path, but exists in:
\\wsl$\ubuntu\home\me
Change the MATLAB current folder or add its folder to the MATLAB path.
When I check the path, it looks like this directory is on the path:
>> path
MATLABPATH
\\wsl$\ubuntu\home\me
I can further verify that this directory is my present directory:
>> pwd
ans =
'\\wsl$\ubuntu\home\me'
and that getEnvFiles.m is in this directory:
>> ls
. .emacs.d HarborData
.. .emacs~ RawHarborData
.bash_history .landscape at
.bash_logout .motd_shown getEnvFiles.m
.bashrc .profile test.m
.bashrc~ .sudo_as_admin_successful
.emacs
Is the issue that I'm using wsl (Windows Subsystem for Linux), or do I have some other misunderstanding?
Type rehash and then try running your script again. Even though you have added the new directory to your path, you need to update the path cache so that it knows about the new scripts it can see.
The problem seems to lie in WSL's ability to add new files to the directory. When I create a new script within MATLAB, and try to run it, I get the problem discussed above. However, running already existing files is not a problem. For now, my only solution is to close MATLAB after creating a new script and reopen it. Then, I can run it. Although, oddly, I can't open it in MATLAB's editor.....
This may or may not help, but ...
Sometimes you have to prepend execution with .\ in order to run scripts in PowerShell or from within the command prompt (in Windows, I'm not sure about other operating systems).
I get this error quite often, especially when not in Admin mode.
Does this work?
% Get path of executing script.
filePath = matlab.desktop.editor.getActiveFilename; % Note that this isn't necessarily the same as the output of 'pwd()'.
% Or: filePath = mfilename('fullpath')
% Make sure all nested directories of filePath are on the path, then tell MATLAB where you're working.
addpath( genpath( filePath ) ); % If this fails, try one folder up the tree:
% addpath( genpath( fullfile( filePath, '.' ) );
cd( filePath );
Alternatively, it looks like you might have written and saved the script, and then typed it into the command window to execute. If it's not a special type of script (Function/Class/GUI/etc.) then you can simply click 'Run' in the Editor tab (or the F5 key) and MATLAB should prompt you with a 'Change Folder' option, to which you should acquiesce.
If you're running this script through the WSL terminal, try my suggested code.
How can I allow user to add additional dependenies after the source code was compiled with mcc.
I was thinking about an empty folder next to the executable, where the users can add the needed .mat-files, but I can't add the folder path to my executable (since addpath is not allowed in deployed applications).
Any ideas?
This answer assumes that your code can be customised by data contained in one or more .mat files at runtime.
You can point your code to look at a folder where the optional .mat file(s) would be located.
For example in the users home folder with a sub folder being the name of your application (or in the local app data) or whereever...
If you want it in a sub folder where the exe is, you can do this as well, you find the exe path using (on windows):
[status, result] = system('path');
installpath = char(regexpi(result, 'Path=(.*?);', 'tokens', 'once'));
fprintf ( 'The exe install path is "%s"\n', installpath );
Then your code looks to load for example:
file2load = fullfile ( installpath, 'subFolder', 'runtimeCustomisation.mat' )
if exist ( file2load, 'file' ) == 2
"doSomething with the file"
end
Or something like that.
Recall that this is for dependencies which are .mat files only.
in my startup.m-file I call a script init.m. In this init file I determine the folder containing this file, want to set the current directory to this folder and continue with some other initialization stuff.
But when I start Matlab, the file will be executed, but the cd command seems to be ignored and the directory won't get changed. Does somebody know why? I.e. the matlabpath variable is extended by the subfolders....
startup.m:
run 'D:\FloatingZone\mscrystalgrowth\trunk\MATLAB\FzKameradaten\FzCameraEvaluation\init.m';
init.m
% get path to folder containing this function
folder=mfilename('fullpath');
file=mfilename();
folder=folder(1:end-length(file));
% change folder
cd(folder);
% add subfolder to the top of the MATLAB path
addpath(genpath('Tools'));
MATLAB did exactly what you told it; there's nothing special about the startup scripts.
From the documentation for the run command:
run changes to the folder that contains the script, executes it, and resets back to the original folder. If the script itself changes folders, then run does not revert to the original folder, unless scriptname changes to the folder in which this script resides.
So your cd command didn't actually do anything, run had already changed the directory. And then changed it back, exactly as expected.
I'm not 100% clear on what you want your startup to do. As already pointed out run is working correctly.
Alternative startup.m
if isdeployed == false
cd 'D:\FloatingZone\mscrystalgrowth\trunk\MATLAB\FzKameradaten\FzCameraEvaluation\'
init();
end
function init.m stored in your FzCameraEvaluation\ folder.
function init
addpath ( genpath ( '../Tools' ) );
end
Does that work?
I see no way to set destination directory or file here: http://www.cs.cmu.edu/~quake/triangle.switch.html
Actually, the program places result file in the same directory, even if current directory is different.
Why? Is ti possible to change?
The output files for the program are generated from the input file names. You can see this from the source code on line 3586
strcpy(b->outnodefilename, b->innodefilename);
...
strcat(b->outnodefilename, ".node");
strcat(b->outelefilename, ".ele");
...
Because of that I don't think there is a way to set the output directory as an option. It seems you will need to manually copy the output files to a different directory
cp output.node your/output/dir/output.node && rm output.node
So i had this issue that occurred when I ran a Matlab script. Here is an a simple example that illustrates it:
So its important to outline the file structure:
MainFolder
script.m
SubFolder
a1.csv
a2.csv
a3.csv
now say i have a script like this:
-> script.m
dir
it would simply print out the files in the folder.
Now the wierd thing, if i run the script in the Subfolder like this:
>>script
it will do this:
>> a1.csv a2.csv a3.csv
but if i do this in the folder:
>>run('C:\Users\....\MainFolder\script.m')
it will only print out
>> script.m
So obviously it is acting as if i ran it form MainFolder rather than SubFolder.
What is the point of this functionality?
The dir command shows the directory contents of Matlab's current directory, not that of where the script is located. So the script showed you the directory contents of wherever you happened to be in the Matlab command prompt when you called that script.
To get what you want, use this in the script:
dir(fileparts(mfilename('fullpath')))
Use pwd to see current dir
Use cd to change directory
Use path to see if your project folders are included in the path
Use which to see you are calling the right *.m file (in case there is multiple .m files with same name on the path)