Why do I receive the error "Attempt to execute SCRIPT FILENAME as a function"?

I receive the error:

 ??? Attempt to execute SCRIPT untitled as a function.

Answer 1

[MathWorks Support Team]

This error results because there is a MATLAB script named [filename] and your calling syntax requires that this script either takes arguments or returns values, which are properties of a function and not of a script. For example, if you have a script called 'banana.m' containing:

disp('banana');

and you call the script with two input arguments as though it were a function by typing the following at the Command Prompt:

banana(1,1)

Then you will get the error. To call the script without error, try calling the script without arguments:

banana

Second, the error can happen because there is a script [filename] and a function [filename] that both exist on the path. If the script appears before the function on the path, then your commands will attempt to use it instead of the function. In Release 2006b (R2006b) and newer versions of MATLAB, the error message will display the location of the script that is called. Rename this script to a more appropriate name. For previous product releases you may use the command:

which <filename>

to determine which instance of [filename] that MATLAB is attempting to call. If you want to see all instances of a function on the path, type the command:

which <filename> -all

Alternatively, such behavior may occur in certain scenarios where your filename contains invalid characters. In particular, you may see this error if the filename contains characters that are MATLAB operators such as -, +, or *. MATLAB filenames must begin with a letter, which may be followed by any combination of letters, digits, and underscores.

[Walter Roberson]

I think you might have accidentally edited the file and added some characters before the "function" keyword

Answer 2

[Lode]

Maybe a stupid remark, but it took me 1h to find...

run(filename.m) doesn't work. You should use run('filename.m') if you didn't put the filename in a variable. This is a very basic mistake but maybe useful to somebody...

Answer 3

[Bruce Lin]

As mentioned in my comment to JMS above, there is a different way you can get this error which the Mathworks respondents did not cover.

If you have a script called ABC.m and a script called ABC-copy.m, and try to run ABC-copy, it will fail with the error above "Attempt to execute SCRIPT FILENAME as a function"

MATLAB interprets the minus sign as an operator, not part of a filename, and the fact that you ALSO have a script called "ABC.m" means that MATLAB tries to execute

"ABC.m - copy.m"

You can prove this by temporarily deleting ABC.m and running 'ABC-copy'. You will see that you get a file not found error instead where MATLAB cannot find ABC.m

One would think that this could be corrected by executing 'ABC-copy.m', i.e. explicitly specifying the .m extension, but it doesn't fix the problem for me.

Hope this helps someone in the future.

[Steven Lord]

Function names in MATLAB must:

• Start with a letter
• Consist only of letters, numbers, and the underscore character
• Not be a keyword (see the output of ISKEYWORD)
• Be no more than NAMELENGTHMAX characters in length.

The function name "ABC-copy" violates the second of those restrictions. I would probably rename that function to "ABC_copy" instead.

[Bruce Lin]

Good point, thanks!

Answer 4

[Walter Roberson]

In addition:

As a practical matter, when basic built-in routines such as disp() are the ones being mentioned as scripts, the problem is almost always that the installation is corrupt for any of several reasons, and reinstallation is required.

A lesser possibility is that the installation is okay but the user has managed to set the MATLAB path to include library directories that should not be named directly, such as if the documentation directory has been added to the MATLAB path. In such cases, restoredefaultpath can be used.