UNIX for Dummies Questions & Answers

Hello experts,

I have a series issue in script that result with bad peformence and I wonder if you can assist me.

For example
I have two files:

File-New, size 15Mb.
File-Old, size 1Mb.

File-New content:
a
b
c
d
File-Old content:
d
f
a
b
c
Result fiel should be:
!a
!b
!c
!d
-f

The current script comapre the lines between these two files in a special order and it takes so long.

The script that currently runs, do the following:
#!/bin/sh
#Input:
# $1 means File-Old
# $2 means File-New
#! - MarkForChange
#- MarkForDeletetion
while read line
do
grep -Fi "$line" $2 >/dev/null 2>&1
if [ "$?" -eq "0" ]; then
echo "! $line"
else
echo "- $line"
fi
done < $1


Do you have any idea how to improve the process.(It take more then hour to run...:-(

Thanks,
Roy.

I must search every line (one of 1000) in 15-20Mb file and mark ! if it was found or - if doesn't.

The script runs very slow.

Any idea?

you might want to look into using 'comm'

And pls don't multipost and start different threads.
Read the rules before posting.

Sorry for thread stuff

I just change the grep -Fi to fgrep and it is fast man...

I still don't undersatnd why is so different on performence...

Shouldn't be any different fgrep is just script that runs grep -F:

#!/bin/sh
exec grep -F ${1+"$@"}

not on Sun/Solaris it ain't!

BTW,

Posix extended grep and the extentions incude -E and -F that duplicate the old egrep and fgrep. On Solaris, this behavior can be obtained by using /usr/xpg4/bin/grep. But read the man page carefully, the old grep has a few things that the new grep doesn't.

And yes, "grep -F" and fgrep should work the same wherever "grep -F" is supported.