I have something like: ------------------------------------------------------------------------ r2 | username | 2011-01-16 16:52:23 +0100 (Sun, 16 Jan 2011) | 1 line Changed paths: D /foo Removed foo ------------------------------------------------------------------------ r1 | username | 2011-01-16 16:51:03 +0100 (Sun, 16 Jan 2011) | 1 line Changed paths: A /foo created foo ------------------------------------------------------------------------ My target is to identify the file added by the "username" in a specific date. Thus, I need to have the combination (username, 16 Jan 2011, A) to insure that it is the right file ands then print foo. My idea is to: delete the white spaces change the newlines into | get rid of the --------------- and replace them with newlines but the problem is that I couldn't replace the ------- since they are mixed with other characters. ------------------------------------------------------------------------ |r2|username|2011-01-1616:52:23+0100(Sun,16Jan2011)|1line|Changedpaths:|D/foo|Removedfoo| ------------------------------------------------------------------------ |r1|username|2011-01-1616:51:03+0100(Sun,16Jan2011)|1line|Changedpaths:|A/foo|createdfoo| ------------------------------------------------------------------------ So I thought it would be a good idea to start by replacing the --------------- by a special character like ||| and then change this character by a newline using awk FS=||| OFS=\n Can anyone help me! thanks