How to run a batch of pairs of files?

I have what I assume is a dumb question, yet I cannot find an answer to it. I want to write a programme to run a bunch of tests through a function and store the answers produced. Each run of the function requires two files, and each file has a pair with which it must be run. Something like: runfile('1a.txt', '2a.txt'); where 1a.txt and 2a.txt are a pair of files. How could I get a program to run each pair of files? I assume I should first separate the two types of files into separate folders. Once I do that, and they are in the same order in each folder (ie. file 1a is the first file in folder 1, and file 2a is the first in folder 2 etc.), what should I use to get the files from the folders and run them through the function?

[Fangjun Jiang]

What makes the two files to be a pair? As long as you can specify the rules (maybe by naming like '1a' paired with'2a', '1b' paired with '2b'?), there should be a way to pair them.

[Charles Dillon]

Yeah, that's what I meant. I meant to say that in the question. 1a goes with 2a. 1b with 2b, etc.

Answer 1

[Jan]

Files do not have an order in folders. Although DIR replies the files in alphabetically sorted order, this is not guaranteed.

[EDITED] You have explained the rule. Then: [EDITED 2] Considering your comments:

function runbatch(sml_folder, iml_folder)

smlDir  = dir(fullfile(sml_folder, '*.sml'));
smlFile = {smlDir.name};
for iFile = 1:length(smlFile)
  aFile = smlFile{iFile};
  aName = fileparts(aFile);
  aSMLFile = fullfile(sml_folder, aFile);
  aIMLFile = fullfile(iml_folder, [aName, '.iml']);
  runfile(aSMLFile, aIMLFile);
end

[Charles Dillon]

Thanks for the help. Unfortunately, this doesn't seem to work. The following is my code:

function [output1, output2]=runbatch(smlfoldername, imlfoldername)

Folder1 = fullfile(tempdir, smlfoldername);

Folder2 = fullfile(tempdir, imlfoldername);

smlDir = dir(fullfile(Folder1, '*.sml'));

smlFile = {smlDir.name};

output1=zeros(length(smlFile));

output2=output1;

for iFile = 1:length(smlFile)

aFile = smlFile{iFile};

aName = fileparts(aFile);

[output1(iFile), output2(iFile)]=runfile(aFile, fullfile(Folder2, [aName, '.iml']));

end

end

(I assumed you'd meant the Folder1 in the last line of your code to be a Folder2, but it doesn't work either way anyway, so that isn't the issue). The code doesn't even run the loop. I rewrote it to output the smlDir and smlFile variables, and got the following:

smlDir =

0x1 struct array with fields:

name

date

bytes

isdir

datenum

smlFile =

{}

[Jan]

The "tempdir" is just the TEMP folder as an example. Folder1 should be smlfoldername and Folder2 should be imlfoldername.

"output1=zeros(length(smlFile))" is a bug, because it creates a square matrix, and not a vector. Use "output1=zeros(1, length(smlFile))" instead. For a pre-allocation create output2 by the ZEROS command also, which is slightly faster than a copy of another array.

[Charles Dillon]

Ok, thank you. So, to clarify, should I have

Folder1 = fullfile(smlfoldername);

Folder2 = fullfile(imlfoldername);

or

Folder1 = fullfile(smlfoldername, 'sml_files');

Folder2 = fullfile(imlfoldername, 'iml_files');

?

Or something else? If it is the second case, what are the purpose of the strings 'sml_files' and 'iml_files'?

[Charles Dillon]

Never mind, I got that part working, thank you!

[Charles Dillon]

It doesn't seem to be finding the files. I'm operating from a folder which contains all my functions and the two folders smlfiles and imlfiles which contain all the files needed, and yet when it runs the function within the loop, which contains usages of dlmread to read the contents of the files passed through by the runfile function, and these dlmreads cannot find the files passed through by the function (presumably because they are only looking in the folder from which I'm operating, rather than the subfolders which contain the files). How could I deal with this issue?

[Jan]

See the newly edited answer above: Use FULLFILE to create absolute file paths.

[Charles Dillon]

Thanks! That works perfectly.

Answer 2

[Stuart Layton]

Why can't you write your test function to take two inputs, one for each file?

[Charles Dillon]

I don't get what you mean. The function can, and indeed, probably must take two inputs, that isn't a problem. I want to run every pair of files, where there are around 20-30 pairs of files. The example of a test function I gave above (runfile('1a.txt','2a.txt') does take two inputs. I just don't know how to get the files from the folders they are in.

[Fangjun Jiang]

help dir

[Stuart Layton]

yep, dir is what you want.

Answer 3

[Fangjun Jiang]

So, what is the pairing rule? Why don't you show us the complete file names in one set and the pairing file names in another set? The following example shows how it can be done if the rules are as simple as you explained so far.

FileOneSet={'1a.txt','1b.txt','1c.txt'};
FileTwoSet=strrep(FileOneSet,'1','2');
for k=1:3
    runfile(FileOneSet{k},FileTwoSet{k});
end

[Charles Dillon]

The pairing rule is that they have the exact same names, with a different extension (one is .sml, the other .iml).

ie. One file will be called bg-38-20.sml (that is the name of one of the files) and its pair will be bg-38-20.iml.

[Fangjun Jiang]

See Jan Simon's solution above. You've got the idea, right?

[Charles Dillon]

I'm not really sure. The problem is that I don't really get from the MATLAB explanations what the outputs of the "dir" and "fullfile" functions are.

[Jan]

DIR creates a list of files from a directory, similar to the DIR command in Windows or ls in Unix, but with a struct vector as output.

FULLFILE concatenates parts of a file path considering initial and final file separators:

fullfile('C:\Temp', 'SubFolder'), fullfile('C:\Temp\', 'SubFolder'), fullfile('C:\Temp\', '\SubFolder'), fullfile('C:\Temp', '\SubFolder') reply the same string 'C:\Temp\SubFolder', such that the programmer do not have to care about the separators.

[Fangjun Jiang]

If you don't understand a particular function, you can type "help dir" or "doc dir" for example to find its usage and syntax. You can also put a break point in your .m file and run step by step to understand each line of code better.