cgrep — a grep enhancement to show preceding lines

Solaris grep has no way to print­ing the lines sur­round­ing a match. Yet my most com­mon use of grep is to locate par­tic­u­lar strings in a log file, with the goal of look­ing at the log mes­sages pre­vi­ous to them. There is a sim­ple solu­tion for this. And also, please note, I did not write any of the code linked here. Just post­ing it to help oth­ers find it.

cgrep shell script: http://examples.oreilly.com/upt2/split/cgrep.sed

#!/bin/sh
#  cgrep - context grep via sed
#  Written by Greg Ubben, DoD, 10 Dec 1993
#
#  Cgrep finds all occurrences of <pattern> in <file...>, showing <context>
#  additional lines of context above and below each occurrence.  <context>
#  must be at least 1, and defaults to 2.  Each new block of context is
#  preceded by the line number of the first occurrence within that block.
#  <pattern> must be a grep-style regular expression, except that you can
#  match across lines, so use "\nWORD" vs "^WORD" and "WORD\n" vs "WORD$".
#
#  The following example will find all occurrences of the word "five" where
#  it is followed by the word "seven" somewhere within the next 3 lines:
#	cgrep -3 "five.*seven"

USAGE="$0 [-context] pattern [file...]"

n=3
case $1 in -[1-9]*)
	n=`expr 1 - "$1"`
	shift
esac
re=${1?}; shift

sed -n "
	1b start
	: top
	\~$re~{
		h; n; p; H; g
		b endif
	}
		N
		: start
		//{ =; p; }
	: endif
	$n,\$D
	b top
" "$@"

This is extra use­ful for pro­duc­tion sys­tems where you can't install pack­ages or make changes to /bin. Just copy the script into your home direc­tory, then either edit your PATH to include your home direc­tory, or ref­er­ence this directly.