CS 497C – Introduction to UNIX
Lecture 29: - Filters Using Regular Expressions – grep and sed
Chin-Chih Chang chang@cs.twsu.edu

Regular Expressions
* egrep’s extended set includes two special characters - + and ?. They are 
  often used in place of * to restrict the matching scope.
* + - matches one or more occurrences of the previous character.
* ? - matches zero or one occurrence of the previous character.
$ egrep "true?man" emp.lst

Regular Expressions
* The |, ( and ) can be used to search for multiple patterns.
$ egrep 'wood(house|cock)' emp.lst
* sed is a multipurpose too which combines the work of several filters.
* Designed by Lee McMahon, it is derived from the ed line editor.
* sed is used to perform noniteractive operations.

sed: The Stream Editor
* sed has numerous features - almost bordering on a programming language but 
  its functions have been taken over by perl.
* Everything in sed is an instruction. An instruction combines an address for
  selecting lines with an action to be taken on them:
sed options 'address action' file(s)
* The address and action are enclosed within single quotes.

sed: The Stream Editor
* The components of a sed instruction are shown as below:
sed '1,$ s/^bold/BOLD/g' foo
     address      action
* You can have multiple instructions in a single sed command, each with its 
  own address and action components.
* Addressing in sed is done in two ways: 
 - By line number (like 3,7p).
 - By specifying a pattern (like /From:/p).

Line Addressing
* In the first form, the address specifies either a single line or a set of 
  two (3,7) to select a group of contiguous lines. 
* The second one uses one or two patterns.
* In either case, the action (p, the print command) is appended to this 
  address.
* You can simulate head -3 by the 3q instruction in which 3 is the address 
  and q is the quit action.

Line Addressing
$ sed '3q' emp.lst
* sed uses the p (print) command to print the output.
$ sed '1,2p' emp.lst
* By default, sed prints all lines on the standard output in addition to the 
  lines affected by the action. So the addressed lines are printed twice.

Line Addressing
* To overcome the problem of printing duplicate lines, you should use the -n 
  option whenever you use the p command.
$ sed -n '1,2p' emp.lst
$ sed -n '$p' emp.lst
* To reverse line selection criteria, use !.
$ sed -n '3,$!p' emp.lst
* To select lines from the middle, do as:
sed -n '9,11p' emp.lst

Line Addressing
* You can select multiple sections as follows:
$ sed -n -e '1,2p' -e '7,9p' -e '$p' emp.lst
* The second form of addressing lets you specify a pattern (or two) rather 
  than line numbers. 
* This is known as context addressing where the pattern has a / on either 
  side.
* You can locate the senders in your mailbox:
sed -n '/From: /p' /var/mail/cs497c

Context Addressing
sed -n '/^From: /p' /var/mail/cs497c
sed -n '/wilco[cx]k*s*/p' emp.lst
* You can also specify a comma-separated pair of context address to select a 
  group of contiguous lines:
sed -n '/johnson/,/lightfoot/p' emp.lst
* To list files which have written permission for the group:
ls -l | sed -n '/^.....w/p'