Bash Scripting Problem...

utopian_penguin · 11-01-2006, 06:00 AM

Dear Scripting Gurus,
I have a problem creating a new script, which gets some specific information from a logs file, the problem is that I want to grep all the lines in the file until a specific pattern is found.

so for example if the file has these lines:

Suse
Fedora
Ubunto
Gentoo
Mandrake
Freespire
Red hat

=================
I want the script to scan the file and print all the lines until it finds Mandrake and stops, which would give in the above example:

Suse
Fedora
Ubunto
Gentoo

Thank you for your help in advance.

druuna · 11-01-2006, 06:03 AM

Hi,

This should work:

sed -n '1,/Gentoo/p' infile

Hope this helps.

utopian_penguin · 11-01-2006, 10:01 AM

It worked, thank you buddy, you dunno how much time you've saved me with this..

Thanks again

I should do more

on awk and sed, they're miraculous!

druuna · 11-01-2006, 10:22 AM

Hi,

You're welcome

Sed and awk are indeed very nice and powerful tools for everyday use and for using in shell scripts.
There's a lot of information on-line for both. If you are interested in printed information take a look at O'Reilly's sed&awk book.

utopian_penguin · 11-01-2006, 01:05 PM

Drunna,
can you explain the options and the syntax of your proposed command.

what are the -e -n used for,and /p!!

druuna · 11-01-2006, 01:48 PM

Hi,

sed -n '1,/Gentoo/p' infile

The -n is needed to suppress the normal output of sed (try not using it), only what is found needs to be printed and that is done by the p at the end.

the 1,/Gentoo/ part is the address range. This states: start at line 1 all the way up to and including Gentoo. The last part must be between forward slashes to indicate that this is a regular expression, not an actual line number. A 'normal' address range looks like this: 3,12 which indicates lines 3 to 12 (including).

The main thing is that you can use regular expression(s) as address indicators. The next example will print all between (and including) Ubunto and Freespire:

sed -n '/Ubunto/,/Freespire/p' infile

The -e option is not used in my example, but you can string sed statements together:

sed -e 's/o/XX/' -e 's/a/YY/' infile is the same as sed 's/o/XX/' infile | sed 's/a/YY/'. The first one being better, it's a lot more resource friendly.

Here's an on-line tutorial for sed.

Hope this clears things up a bit.

DotHQ · 11-01-2006, 02:14 PM

Thanks for the explanation, and good link to the tutorial!

druuna · 11-01-2006, 02:29 PM

Always nice to know that others find it helpful too.

utopian_penguin · 11-07-2006, 03:46 PM

Hi Drunna,
How are you buddy, here's another challenge I found creating my script, my log file contains Timestamps, what I want to do is to grep a specific number of lines around the - for example - Gentoo line,

In another way, I want to have the chunk of lines around the line I want (which is included of course) , a period of time in my case ...

Please, Help

druuna · 11-08-2006, 12:55 AM

Hi,

This is indeed a challenge, but no fear sed is here

Code:

#!/bin/bash
# -------------------------------------------------------------------------- #
# prints out 4 lines around pattern
# -------------------------------------------------------------------------- #

case $# in
  2);;
  *|1) echo "Usage: $0 pattern <infile>";exit;;
esac;

infile=$2

sed -n '
'/$1/' !{
  # does not match - add this line to the hold space
  H
  # bring it back into the pattern space
  x
  # Two lines would look like .*\n.*
  # Three lines look like .*\n.*\n.*
  # Delete extra lines - keep two
  s/^.*\n\(.*\n.*\)$/\1/
  # now put the two lines (at most) into
  # the hold buffer again
  x
}
'/$1/' {
  # matches - append the current line
  H
  # get the next line
  n
  # matches - append the current line
  H
  # get the next line
  n
  # append that one also
  H
  # bring it back, but keep the current line in
  # the hold buffer. This is the line after the pattern,
  # and we want to place it in hold in case the next line
  # has the desired pattern
  x
  # print the all lines
  p
  # add the mark
  a\
---
}' $infile

# -------------------------------------------------------------------------- #
# End

There will be a mark (---) printed at the end of each hit. Just remove the a\ and the --- from the script if that is not needed.

A sample run:

Code:

./blaat Gentoo infile
Fedora
Ubunto
Gentoo
Mandrake
FreeSpire
---

I included as much comments as possible. Here's the same script, without all the comments:

Code:

#!/bin/bash
# -------------------------------------------------------------------------- #
# Prints out 4 lines around pattern.
# -------------------------------------------------------------------------- #

case $# in
  2);;
  *|1) echo "Usage: $0 pattern <infile>";exit;;
esac;

infile=$2

sed -n '
'/$1/' !{
  H
  x
  H
  x
  s/^.*\n\(.*\n.*\)$/\1/
  x
}
'/$1/' {
  H
  n
  H
  x
  p
  a\
---
}' $infile

# -------------------------------------------------------------------------- #
# End

Hope this helps.

utopian_penguin · 11-08-2006, 01:31 PM

CooooooooooooL , You Rock dude

, you always have the answers present

!!!!!!!!!!!

Thanks a zillion

druuna · 11-08-2006, 01:39 PM

Hi,

You're welcome