Hi, How to remove ^M character from a file. A sample script would be helpful thanks sharath

Hi, in vi do the following: :g/^M/s/// (the ctrl-m has to be entered with the key strokes "ctrl-v followed by ctrl-m". That's teh only way it works! Or use sed to do it: sed 's(ctrl v ctrl m)g//g' old.file > new.file Bye!

And with 'tr' : tr -d '\r' <old.file >new.file Jean-Pierre.

Hi: You can also use the octal representatiion of ^M. The following gets rid of control-M as well as control-Z (DOS eof marker): tr -d '\015\032' < file You can also embed these sequences in sed: sed 's/'"$(printf '\015')"'$// s/'"$(printf '\032')"'$//' file Regards, Nails

One other way is to use the built-in command. I am not sure if HP-UX has this command, but with Solaris you can just do: dos2unix <dos file> <unix file> You can also go back with unix2dos.

HPUX has the command "dos2ux" instead of dos2unix. Save the following as a script "dos2unix" and it will work the same as other systems. #!/bin/sh tmpfile="/tmp/$$" if [ $# -ne 1 ] ; then echo "USAGE $0 dosfile" exit 4 fi dos2ux $1 > $tmpfile && mv $tmpfile $1

The best way is to avoid getting them in the file in the first place. The most common way this happens is that you ftp the file with the wrong option of Binary or Ascii. If you use Binary on an FTP it does nothing to the file, but if you use Ascii it attempts a conversion to UNIX, or DOS format. For example if you often ftp windows text files to unix you should use the Ascii mode for ftp so that it will strip off the ^M. If you don't know which option to use, just try one option, and if it doesn't work, then just retransfer the file using the other option and it should be fine.