Question

Possible Duplicate:
Regex – nested patterns – within outer pattern but exclude inner pattern

I am trying to get a substring of a string after/from a word. But I want that word to be outside of the parenthesis. For example:

something (theword other things) theword some more stuff should give me theword some more stuff instead of theword other things) theword more stuff. How can I do this in regular expressions. I am using PCRE (e.g. php, python regex engine)

Edit:

The string I am trying to use this regular expression on is a mysql statement. I am trying to remove parts until FROM part, but inner sql statements (that are in parenthesis causing problems to me).

Answer 1

This is a task similar to the one described in this question: Regex – nested patterns – within outer pattern but exclude inner pattern

See my answer for why it’s not 100% possible and for a hack-ish solution that works most of the times (for shallow nesting, that is).

Update:

If you know that parentheses wil neither be nested, nor escaped, you could use something akin to this:

(?<=\)|^)([^()]+)(?=\(|$)

So with this as the haystack:

something (theword other things) theword (some) more stuff

Wou would end up with these snippets in capture group 1:

something 
 theword 
 more stuff

Note that there are leading/trailing spaces captured as well, to prevent them use this instead:

(?<=\)|^)\s*([^()]+)\s*(?=\(|$)

Same regex pattern, but with comments:

(?<=\)|^) #either ")" or beginning of string
\s* #optional leading whitespace
([^()]+) #any sequence of characters but "(" or ")"
\s* #optional trailing whitespace
(?=\(|$) #either "(" or end of string