I have this text
XXX text XXX XXX text XXX XXX text XXX and i want to capture the text between the XXX and XXX. (i am trying to get chapters out from a book )
/XXX.*XXX/ This will capture the first begin and the last end
/XXX.*?XXX/ This will skip every second chapter
Thanks ahead Barak
If you text contains line feeds (\n) you'll need to add the "dot matched newline" switch to your regex, as well as making your match "non greedy":
/(?s)XXX.*?XXX/ Edited: Thanks to Alan's comment - I had the wrong switch: (?s) is correct
^ and $, so it would be useless here. Perhaps you're thinking of "single-line" or "DOTALL" mode, which enables the . to match linefeeds.
Solution using sed
$ sed -n '/XXX/,/XXX/{n;p}' text text text text If this XXX strings are always in separate lines, i would suggest simple iterating through lines and picking it 'by hand'. It should be faster than multi-line regexp.
python :
delim = "XXX" inside = False lines = [] idx = 0 for line in file: if line.strip() == delim: inside = not inside if inside: lines.append([]) else: idx += 1 elif inside: lines[idx].append(line) Your description doesn't really match your examples. If XXX is supposed to represent a chapter heading, there would only be one at the beginning of each chapter. To detect the end of a chapter, you would need to do a lookahead for the next chapter heading:
/XXX.*?(?=XXX)/s That should work for all but the last chapter; to match that you can use \z, the end anchor:
/XXX.*?(?=XXX|\z)/s It really would help if we knew which regex flavor you're using. For example, in Ruby you would have to use /m instead of /s to allow . to match linefeeds.