grep for string but include lines above and below

[Derek Brown]

Hi, This forum has always been great for me and I wonder if anyone can solve this one for me please ?

I need to search a file for a specific string. Easy enough with grep e.g.

# grep root /etc/passwd

BUT .... I need the answer to provide the 3 lines above the string and also the 2 lines below the string.

Linux has this built in as a feature but is the same thing possible in standard HPUX ?

Many thanks.

-------------------------------

P.S. This Thread has been moved from HP-UX --> System Administration to HP-UX --> Languages & Scripting - Forum Moderator

[Pete Randall]

Derek,

From "Handy One-Liners for Sed" (attached):

# print 1 line of context before and after regexp, with line number
# indicating where the regexp occurred (similar to "grep -A1 -B1")
sed -n -e '/regexp/{=;x;1!p;g;$!N;p;D;}' -e h


Pete

[Peter Godron]

Derek,
please have a look at:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=861863
Seems for multiple lines you require gnu grep.

[john korterman · Accepted solution]

Hi,

try the attached script, find_bef_after.sh
using $1=string to grep for
$2=number of lines before and after string
$3=path to file

slow, but worked the last time I tried..

regards,
John K

[Derek Brown]

To John Korterman

John .... did you omit to attach the sample script find_bef_after.sh ?

Thanks

[john korterman]

sorry, hope it appears now..

[James R. Ferguson]

Hi Derek:

How about a perl script like:

# cat ngrep
#!/usr/bin/perl
#@(#)ngrep $ Show +-3 lines matching target January 2006 JR_Ferguson $

use strict;
die "Usage: $0 pattern file\n" unless $#ARGV == 1;

my $pat = shift;
my $file = shift;
my @lines;
my $limit;
my ($i, $m, $n);

open( FH, "<", $file ) or die "Can't open $file : $!\n";
@lines = ;
$limit = $#lines;

foreach (@lines) {
if ( m/$pat/ ) {
$m = $i-3 < 0 ? 0 : $i-3;
$n = $i+3 > $limit ? $limit : $i+3;
for ( $m; $m <= $n; $m++ ) {
print $lines [$m];
}
print "\n";
}
$i++;
}
1;

Regards!

...JRF...

[john korterman]

Hi again,

a variation taking four parameters - if it ever gets there:
$1 the string to match
$2 the number of lines to print before the matched string
$3 the number of lines to print after the matched string
$4 the file itself


regards,
John K.

[Hein van den Heuvel]

JRF's solution will work fine a a file like /etc/passwd, but as it sucks in the whole file it will become problematic for very large (100+MB?) files.

I prefer a small circular buffer to just retain the last few lines.

For example:

awk 'BEGIN {W=4} {line[i++%W]=$0} /targetstring/{for (j=i-W;j
Or with a before-and-after:
awk 'BEGIN {W=5} {line[i++%W]=$0; if (1==x--){for (j=i-W;j
de-composed:
BEGIN set window size
Main loop
1) move current record into circular buffer
2) if counter x equals 1, then target was seen 1/2 a window lines ago, start print loop of memory array.
3) /target/... if target is seen, then set counter to 1/2 the window size.

( This code does NOT handle mutlipel adjacent matches )

I also have posted a perl script in:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=630972

Other prior discussions in:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=524659

Or a little further back in time:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=224110


btw...
Under OpenVMS it would be:
$SEARCH/WINDO=(before,after) file text


Cheers,
Hein.

[Sandman!]

Derek,

It can be handled easily with ex as:

# ex -s input_file <> //-3,//+2 p
> EOF

cheers!