find a specific file

Hello,

I have different data on a similar numerical model stored in text with similar but different names.

What I want to know is if it possible to find a file base on its name Basically I open these files in my code:

Stresses-UP-7_5-22_5-0MPa.txt

Stresses-UP-8-22-14MPa.txt

Within my code, I would like to find a specific value stored in another text file names:

Stresses_AxSym_UP_7-5_22-5.txt

Stresses_AxSym_UP_8_22.txt

both text files correspond to the same models having dimensions: 1) 7.5 and 22.5; and 2) 8 and 22.

the text file names are different, in one the names are separated by underscore and the other by the minus sign. Would it be possible to code something to find the Stress_AxSym text file based on the information of Stress_UP text file?

thank you

[Cedric]

So you know the first file name and you want to generate the corresponding second file name based on the first? Or are you just asking how to generate all file names based on 7.5, 22.5, 8, 22? In any case, what happens to the 'OMPa' and '14MPa' parts of the first names? Are they discarded?

[Nicolas]

I'd like to use the information in the name of file 1 to find file 2

Answer 1

[Cedric]

You could use the following function:

 function fName_out = convertFilename( fName_in )
    match = regexp( fName_in, '[\d_]+', 'match' ) ;
    dims  = strrep( match(1:2), '_', '-' ) ; 
    fName_out = sprintf( 'Stresses_AxSym_UP_%s_%s.txt', dims{1}, dims{2} ) ;
 end

With that ..

 >> convertFilename( 'Stresses-UP-7_5-22_5-0MPa.txt' )
 ans =
      Stresses_AxSym_UP_7-5_22-5.txt

 >> convertFilename( 'Stresses-UP-8-22-14MPa.txt' )
 ans =
      Stresses_AxSym_UP_8_22.txt

[Nicolas]

That's great. There are lot of functions I didn't know there. I'll play with that for now. Thanks a lot.

[Cedric]

You're welcome.

The call to REGEXP returns sub-strings like '7_5', '8', etc. The advantage over SSCANF is that it is more flexible and we can manage cases where there is a decimal part (7.5) as well as integer cases (8) in one expression. SSCANF doesn't have this flexibility, yet it works much faster than REGEXP when you can use it. If time were critical, I'd build a solution based on SSCANF (as mentioned by ImageAnalyst), but here it isn't, as reading/processing files is orders of magnitude slower than the REGEXP call.

The first two elements of the cell array outputted by REGEXP are the two numbers with the '_' in place of the decimal point. Using STRREP, we replace the underscore (if present) in both with a dash. The generates for example the cell array {'7-5', '22-5'} for the first file name.

We finally print these two strings at relevant locations (where '%s' are) in the pattern describing your output file name format.

Answer 2

[Image Analyst]

You can use sprintf() to create filenames based on variables:

filename = sprintf('Stresses_AxSym_UP_%d_%d.txt', integer1, integer2);

[Nicolas]

nice, thanks. Would it be possible to extract the %d in order to match them in another file.

For example: file 1 named -7_5-22_5- corresponds to the file named _7-5_22_5; since both correspond the the same model with characteristics (7.5 and 22.5) expect that they are written differently.

[Image Analyst]

Sure. Use sscanf().