From: hadi motamedi <motamedi24 at gmail.com> > On 11/30/10, John Doe <jdmls at yahoo.com> wrote: > > The trick is that your original file has '\r' chars lurking around... > > Forgot I did removed them manualy when I saw them... > > cat Edit3 | tr -d "\r" | while read LINE; do > > Instead of just copy/pasting, try to understand what it does. > > Here is how it works: > > - Read each line in the LINE variable. > > - If the line contains the string '>..', it is a "section" line. > > Set the log filename to the section title. > > - If not, just write the line to the current log filename. > Sorry. I didn't get the point clearly. What I need is to separate the > log files from each of the modules. For example, I need all of the log > files coming from XAPP module. In the main log, it can be > distinguished by searching for the following line: > XAPP >.. > Then you see subsequent lines that are logs coming from this module. > Then the logs from another module will come following the previous > one. I think you code is not doing this . Can you please correct me on > my understanding of your code body? My pseudo-code does exactly what you described (re-read it)... And it creates <MODULE>.log files... $ ll -n total 36 -rw-r--r-- 1 2000 500 4678 nov 30 10:49 Edit3 -rw-r--r-- 1 2000 500 39 nov 30 10:54 HLR.log -rw-r--r-- 1 2000 500 2320 nov 30 10:54 IPTR.log -rw-r--r-- 1 2000 500 478 nov 30 10:54 SCCP.log -rw-r--r-- 1 2000 500 754 nov 30 10:54 SNMP.log -rw-r--r-- 1 2000 500 507 nov 30 10:54 TCAP.log -rw-r--r-- 1 2000 500 281 nov 30 10:54 XAPP.log -rwxr-xr-x 1 2000 500 207 nov 30 10:54 test.sh* If you think it is not doing it, I cannot help you... Either ask your sysadmin to help you, or use someone else code... JD