Regular Expression Reference on a Mouse Mat * Training
POSIX details * Perl details * Java details

Grep and egrep are Unix and Linux utilities which filter an incoming file, outputting all lines which match a regular expression and not outputting lines that fail to match. The original grep uses a somewhat restricted regular expression set, and egrep (in full, the extended global regular expression processor) has somewhat further facilities. Fgrep (the f stands for fake) doesn't use regular expressions at all - it checks for literal strings.

Example:

      grep cat demo.txt

outputs all lines from the file demo.txt that include "cat"

Special characters in regular expressions such as [ ] ? and * also have a special significance to the Unix or Linux shell, and it's usual to write your regular expression in quotes to protect it from the shell.

Operator Type	Examples	Description
Literal Characters Match a character exactly	a A y 6 % @	Letters, digits and many special characters match exactly
	\$ \^ \+ \\ \?	Precede other special characters with a \ to cancel their regex special meaning
	\n \t \r	Literal new line, tab, return
Anchors and assertions	^	Starts with
	$	Ends with
Character groups any 1 character from the group	[aAeEiou]	any character listed from [ to ]
	[^aAeEiou]	any character except aAeEio or u
	[a-fA-F0-9]	any hex character (0 to 9 or a to f)
	.	any character at all
Counts apply to previous element	+	1 or more ("some") egrep only
	*	0 or more ("perhaps some")
	?	0 or 1 ("perhaps a")
Alternation	|	either, or egrep only
Grouping	( )	group for count and save to variable egrep only

Additionally, some versions of egrep (especially linux) support numerous extra sequences such as POSIX character groups (e.g. [[:alnum:]] for an alphanumeric character) and extended counts (e.g. {4,8} for bewteen 4 and 8 of the previous element). Take care when using these extras if you want your scripts to be portable.

grep, egrep and fgrep all take a number of options to allow you further control over your matching and output. The options don't differ between egrep, fgrep and grep but there are more options available under linux than there are under Unix. The following table includes the more common options under both operating systems.

option	Description
-i	Ignore case in matching
-v	Output only lines which do NOT match
-c	Output only a count of the numer of lines that match
-l	Output only the names of input files that include at least one match
-n	Output the line number within the file in from of each matched line
-h	Supress file name output even if there are multiple input files

The -r option (not available on all versions of grep) allows the recursive searching of all files in subdirectories.

Back to Regular Expression Home Page
Jump to Elements of a regular expression
Order a Regular Expression Mousemat for £4.95 inclusive