extract number out of the file title for further usage

hi everyone,

i have many files inside a folder. The files have names like 12E.dat, 13.dat, ...100.dat i use the following code to load them

files = dir('*.dat');
for k = 1:numel(files)
D = load(files(k).name);
tof = D(:,2)*1000;
...
end

I want to creat an array like E = [12, 13, ..., 100] taken out of the title of the file. Is their a way to realize this?

thanks a lot in advance

Answer 1

[Jos (10584)]

files = dir('*.dat'); 
names = {files.name} ; % put the names in a cell array of strings
names = strrep(names,'.dat','') ; % remove extension

Names is now a cell array of strings. If they are all numbers, you can convert them using str2double

values = str2double(names)

Answer 2

[Image Analyst]

Did you try sscanf()?

Answer 3

[eFKa]

Thanks to both of you guys. For my purpose suggestion of Jos worked very well.

Have a nice evening.

[Image Analyst]

OK, just be aware that if your filenames are "12E.dat, 13.dat, ...100.dat" like you said, you'll get a nan for the 12E file (or any other file that is not strictly pure numbers", so that's why I recommended sscanf() which is more robust, see:

% Jos code:
values = str2double(filename)
% Image Analyst code:
theValue = sscanf(filename, '%d', 1)

In the command window:

values =
   NaN
theValue =
    12

Notice my code gives the desired 12 for the "12E" filename whereas str2double() gives a nan. So if you want to skip filenames with non-numerical characters in them, you'll have to use isnan() to check for that, or else use my code to pull out just the number. But that code only handles filenames if they start with a number. It could be made even more robust by getting rid of non-numbers wherever they appear. So look at this more robust code:

filename = 'xyz12E.dat'; % Number in the middle
% Jos code:
values = str2double(filename)
% Image Analyst code:
numbersIndexes = (filename >= '0' & filename <= '9') 
filename(~numbersIndexes) = ' '
theValue = sscanf(filename, '%d', 1)

values =
   NaN
numbersIndexes =
     0     0     0     1     1     0     0     0     0     0
filename =
   12     
theValue =
    12

I've just learned to always write super robust code because if I don't, and I roll out my code worldwide and one of my users does something unexpected, they'll blame the software, not themselves, and my software (and me!) will get a reputation as "buggy" or unreliable, and I can't afford that.

[Jos (10584)]

I second that, ImageAnalyst. Note that two distinct files, like 12E.dat and 12F.dat, end up having the same label …

[Image Analyst]

Very good point (that I overlooked). Sometimes things like that will slip by and you'll only realize it when you get an angry email from a user about how your code does not work. Like they say in software development "It's hard to make code idiotproof because idiots are so ingenious." But perhaps the 12E was just a typo - who knows? But my first assumption is that the user meant what he said.

[eFKa]

my files are like 7E.dat to XXE.dat, E indicating energy in Electron-volts. Using the following worked smoothly for me.

 names = strrep(names,'E.dat','') ;

scanf() also seems to be interesting because of its versatility. Thanks for the insight once again