My favourite RegEx

[Marcus Tettmar]

Suddenly dawned on me that there's one regular expression I use almost every day:

(?<=TOKEN1).*?(?=TOKEN2)

What this does is pull all text out from between two other pieces of text matching token1 and token2. Very useful for parsing data from websites.

I should blog this with a website example.

What's your favourite regex?

[Dorian (MJT support)]

I'd never used regex until I saw this post, and it prompted me to do a little research.

It's very powerful. I used it to strip out all the URLs from a text file. Here it is, in case it helps anyone.

Code: Select all

//sample text
let>text=rfuhroiurnfifnroi http://www.fish.com fuh3ifuh34ifurf http://www.chips.com

// it seems pattern looks for http://, not www. so we'll add in the http://
stringreplace>text,www,http://www,text

// Find URLS
RegEx>[Hyperlink],text,1,matches,num,0

// write it all to a file
Let>k=0
Repeat>k
Let>k=k+1
WriteLn>%USERDOCUMENTS_DIR%\url output.txt,result,matches_%k%
Until>k,num

[jpalic]

Marcus,

I use something similar to yours although I never quite understood the meaning of the zero-width positive lookbehind (?<=regex) or the zero-width positive lookahead (?=regex) modifiers.

What issues do you avoid by using those modifiers in this regex?

Jim

PS - There is a good regex reference here:

http://www.regular-expressions.info/refadv.html

[Marcus Tettmar]

What these do is exclude those tokens from the match so that we get only the value between them.