Re: automated text file reading and array element extraction

[Leo Butler]

On Thu, 2 Dec 2010, Michael Hadrien wrote:

< Hello eveyone, 
< I'm a beginner in Octave, running Octave 3.2.4 on Ubuntu 10.10. 
< What I'm trying to do is read numerical data in sequentially named text files 
in two different directories, extract the second column from each file in both
< directories, subtract all the columns in one directory from all the columns 
in the other directory, then create a new directory with a set of files 
containing the
< first column (from any directory, they're constants) and the resulting 
columns from the subtraction. 
< 
< So Directory A has:
< aaa_001.txt
< aaa_002.txt
< aaa_003.txt
< 
< Directory B has: 
< bbb_001.txt
< bbb_002.txt
< bbb_003.txt
< 
< The files all look like this:
< 33  32
< 48  45
< 76  34
< 46  54
< 
< The first column never changes. 
< I want to take second columns from one directory and subtract them from all 
the second columns in the other directory. Both directories have the same 
number of
< files. 
< 
< This is the code I currently have:
< 
< files=dir('*.txt'); #creates struct with all files
< filesarray = struct2cell(files); #converts struct array to cell array
< names=filesarray(1,:) #extracts first row of cell array (where all the file 
names) and assigns to names
< 
< for i=1:length(names)
< eval(['load ' files(i).name ' -ascii'])
< end
< 
< 
< I've repeated this for both directories. Now all the files have been loaded 
and I can extract column 2 from each file individually using:
< 
< aaa_001(:,2)
< 
< But I want to be able to do this for all the files automatically and do the 
subtraction operations. 
< I'm not sure what to search or how to go about this. I've thought of maybe 
creating two matrices with all the col 2 data from each directory on a 
different matrix,
< but I don't know how to do that. Then I'd be able to subtract the matrices. 
The next step would be extracting and writing each column to a new text file.
< 
< Any help I can get would be greatly appreciated!
 
 A general rule: if you are using eval, you are doing it wrong.
 I would try something like:

 filesA=dir('A/*.txt');
 filesB=dir('B/*.txt');
 for i=1:length(filesA)
  filesA(i).data=load(fileA(i).name);
  filesB(i).data=load(fileB(i).name);
 endfor
 
 Now you have your data loaded in memory and you can do what you
 like. Since the data and filenames are associated, you can
 create a new filename and write your transformed data to that
 file in a mechanical way without resort to eval.

 Leo
-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.