Shell Programming and Scripting

Try:

Code:

perl -pe 's/\(+.*?\)+//g; s/([*"])+.*?\1+//g' file

This is using perl regex's lazy matching operator, .*? means a lazy (minimal) match of characters..

The assymetrical enclosing characters use a separate substitution command for each pair. In this case there are only parentheses. In the substitute command they are preceded by backslashes to use their literal meaning.

For the symmetrical enclosing characters in the second substution command, a backreference \1 is used to match the previously matched attribute from the enclosing characters class, in this case either * or ", but any number of characters can be added inside the square brackets. The character class is within parentheses (...)so it can be backreferenced later.


It can also be written more compactly:

Code:

perl -pe 's/\(+.*?\)+|([*"])+.*?\1+//g' file

using alternation (|)


If it is only these three cases then perhaps this is preferable:

Code:

perl -pe 's/\(+.*?\)+//g; s/\*+.*?\*+//g; s/"+.*?"+//g' file

Since it is somewhat easier to read..


Output:

Code:

Hello  Outside bracket 
          Hello (Open Bracket without closed
          Hello  
          Hello  
          Hello

--
For multiline sequences you could use:perl -00pe ....., or perl -0077pe .....so for example:

Code:

perl -00pe 's/\(+.*?\)+|([*"])+.*?\1+//g' file

With a sed command like:

Code:

sed -e 's/[(][^)]*[)]//g' -e 's/[*][^*]*[*]//g' -e 's/["][^"]*["]//g' file

and file containing:

Code:

Input : Hello (Within Bracket) Outside bracket ( within bracket)
          Hello (Open Bracket without closed
          Hello ****WORLD** 
          Hello *WORLD* 
          Hello "World"
*between*after asterisk(between)after parens"between"after quotes
Note that parens are matched 1st, asterisks are matched 2nd, and quotes
are matched last...
(*")"*"(*("*("

you get the output:

Code:

Input : Hello  Outside bracket 
          Hello (Open Bracket without closed
          Hello WORLD 
          Hello  
          Hello 
after asteriskafter parensafter quotes
Note that parens are matched 1st, asterisks are matched 2nd, and quotes
are matched last...
*("

Is this closer to what you're trying to do?