Sed syntax

This syntax format is described in standard "Unix sed utility in IEEE Std 1003.1-2001, Portable Operating SystemInterface (POSIX ), Shells and Utilities".

Special match symbols

^
 begin of string 
$
 end of string 
.
 any symbol 
*
 previous value is repeated 0 or more times 
+
 previous value is repeated 1 or more times 
{N}
 previous value is repeated exactly N times 
{N,M}
 previous value is repeated from N to M times 
[[:digit:]]
 any digit

To use specials symbols you need to screen them with '\'.
For example, '\+'.

Special replacement symbols

&
 whole matched string 
\1
 1 matched atom

Examples

search replace result 
^011.*
 
777
 string 011XXXXX will be replaced with string 777 
^011
 
7
 string 011XXXXXXX will be replaced with 7XXXXXXX 
.*
 
00&
 to all strings will be added a prefix 00.
& - means "put original string"		 
^011
 a prefix 011 will be cut 
^(8)(.{10})$
 
7\2
 strings 8ХХХХХХХХХХ will be replaced with 7ХХХХХХХХХ.
All strings, which contain less or more 11 digits won't match this rule.
Original string is split into 2 "atoms" - digit 8 and 10 remaining digits.
\2 - means, that we put 2nd "atom" - 10 digits after 1st digit 8		 
(^.*)(.{2}$)
 
749511111\2
 system remembers 2 last digits of original string and they are added to string in field "replace".
Therefore, if original string was 444455, the resulting string will be 74951111155		 
.*
 
&
 any value will pass without changes.
& - means "put original string"		 
^[[:digit:]]*$
 
&
 digital string will pass without changes.
& - means "put original string"		 
^\+
 prefix '+' will be cut.
'\' - is a symbol of screening of a special symbol '+'

Русский перевод