Re: Is there something quicker than textscan?

[CdeMills]

This is one of the issues I face in my dataframe package. The option:
separate the processing in two parts, one m-file which does batch
processing, the second one to do individual processing. 

For the batch processing:

for indi = (1:length(varargin))
   
   list_files = dir (varargin{indi});
   
   for indf = (1:length(list_files))
       df = dataframe(list_files(indf).name);, "seeked", \
                      "time", "sep", "\t", "locales", "fr_BE");
       save ("-binary", [list_files(indf).name ".df"],  "df");
   endfor
endfor

This way, the first pass reads every file matching some pattern you give,
and all results are saved as binary. The second pass is about processing
file by file, and reading the binary file is much quicker.

Regards

Pascal

--
View this message in context: 
http://octave.1599824.n4.nabble.com/Is-there-something-quicker-than-textscan-tp4483991p4484291.html
Sent from the Octave - General mailing list archive at Nabble.com.