HI, i need to remove about a 3000000 lines in a "too long file", without using vi editor. Im using hp-ux 11.0.

look into using either one of the standard Unix utilities: sed, ex, awk etc....
if you elaborate on the removal criteria, you might get a more detailed answer.

ex - data << end-of-script
# give your removiing creteria over here
#this will delete line number 2

2d
wq
end-of-script
this deletes the second line

found this on unix guru tip. maybe u could try it. gd luck

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

UNIX GURU UNIVERSE
UNIX HOT TIP

Unix Tip 2429 - August 26, 2005

http://www.ugu.com/sui/ugu/show?tip.today

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=


TRIMMING THE LOG


Various unix processes can
produce fast growing logs that
sometimes need to be trimmed
instead of deleted, for reference
or troubleshooting. And you
likely have no desire to edit the
files. Here is a handy ksh
script that will quickly trim
the log so it keeps recent
information and lets you keep as
many lines as you think you might
need. I call it trimlog:

#! /bin/sh
# trimlog
filesize=`cat $1|wc -l`
trim=`expr $filesize - $2`
if [ $trim -gt 0 ]
then
sed "1,$trim d" $1 > /tmp/$1
mv /tmp/$1 $1
echo $1 trimmed by $trim lines
fi

Use it by feeding in the name of
the log you want to trim followed
by the number of lines you want to
keep:

# trimlog oracle_listener.log 10000


Of course you can't use it in
the /tmp directory. There are
probably newer and more efficient
ways to do this, but this works.

I am brand new to system administration, being a programmer but have no unix experience. Could you please let me know how to perform this same function but instead of giving arguments at command line to do in aprogram so I could put in as a cron job. Greatly appreciate any responses.

Here is something I have found. Is there any way someone could explain to me in 'english' what the code after 'sed' is doing? Also what doesn maxdepth does, have an error that maxdepth is a bad option.

cd /var/sam/log || exit 1

for i in $(find . -type f -maxdepth 1)
do
filename=$(echo $1 | sed 's?^\./\(.*\)?\1?')
cp -p $i old/$(date +'%Y%m%d').$filename
> $i
done
exit 0