From: Gavin Kistner <gavin@...>
Date: 2004-03-12T11:22:49+09:00
Subject: Re: Regular Expression for D(elimiter) Separated Values  File

On Mar 11, 2004, at 3:40 PM, Warren Brown wrote:
> Now to your original intent, it sounds like what you want is to split 
> on any semicolon that is not preceded by a backslash.  This is a 
> little more difficult, since there is no "look-behind" construct in 
> regular expressions.

To be pedantic - not in /Ruby's current implementation of/ regular 
expressions. This should all change in 2.0 with Oniguruma, right?

Anyhow, if the above is an accurate assessment of the needs, just for 
the fun you can do it another way than the (effective) solution already 
provided.

str = "aaa;bbb;ccc\\;ccc;ddd\\\\;eee;fff\\\\\\;fff;ggg\\\\\\\\;hhh";
chunks =  str.gsub(/([^\\])(\\\\)*;/,'\1\2�').split('�').inspect;

The premise here is to replace the desire for a negative lookbehind 
with a consumed character, which is then stuck back into the string 
along with a new 'magic' character which is guaranteed not to be in the 
source string. This character is then used to split the output.

The regular expression above says "find a character which isn't a 
backslash, followed be an even number of backslashes, followed by a 
semi-colon (which we now know must be a field delimiter, since it can't 
have been escaped)".

Seems to work from the naive sample I posted above...but requires a 
magic character to be available.
--
(-, /\ \/ / /\/