This question is more tricky than it really appears. The reason is that you can have months, days, and hours I guess, on one or two digits, and I assume that the number of data points can vary.
The standard approach is probably to use FGETL in a loop, test each line to determine whether it's a timestamp of a data point (it could be based on LENGTH if numbers are really stored using less than 12 chars), etc. Here is an example:
fin = fopen('dataLuis.txt', 'r') ; % Open input file (read mode).
fout = 0 ;
while ~feof(fin)
line = fgetl(fin) ; % Read line from input file.
if isempty(line), continue ; end % Discard if empty line.
if length(line) > 12
if fout ~= 0, fclose(fout) ; end % Close previous file if open.
fname = sprintf('%s.txt', line) ; % Build output file name.
fname = strrep(fname, '/', '-') ; % .. replace / and : that are
fname = strrep(fname, ':', '') ; % invalid in filenames.
fout = fopen(fname, 'w') ; % Open output file (write mode).
elseif fout > 0
fprintf(fout, '%s\r\n', line) ; % Output data line.
end
end
if fout > 0, fclose(fout) ; end % Close open files.
fclose(fin) ;
A less standard approach is based on REGEXP, using a pattern that identifies timestamps and splits the content around matches. Here is an example:
buffer = fileread('dataLuis.txt') ;
[m, s] = regexp(buffer, '[\w/,]*:..', 'match', 'split') ;
for k = 1 : numel(m)
fname = sprintf('%s.txt', m{k}) ;
fname = strrep(fname, '/', '-') ;
fname = strrep(fname, ':', '') ;
fout = fopen(fname, 'w') ;
fwrite(fout, strtrim(s{k+1})) ;
fclose(fout) ;
end
A more elaborate approach still based on REGEXP splits the file content into timestamps and data. Here is an example:
buffer = fileread('dataLuis.txt') ;
pattern = '(?<timestamp>[\w/,]*:..)\s*(?<data>.*?)(?=($|\s\d{1,2}/))' ;
content = regexp(buffer, pattern, 'names') ;
for k = 1 : numel(content)
fname = sprintf('%s.txt', content(k).timestamp) ;
fname = strrep(fname, '/', '-') ;
fname = strrep(fname, ':', '') ;
fout = fopen(fname, 'w') ;
fwrite(fout, content(k).data) ;
fclose(fout) ;
end
