From: Ken Bloom <kbloom@...>
Date: 2006-11-13T23:50:05+09:00
Subject: Re: regular expression too big

On Mon, 13 Nov 2006 00:07:52 -0800, Paul Lutus wrote:

> Peter Schrammel wrote:
> 
>>> When you post an question like this, it is always a good idea to
>> reveal the
>>> purpose as well as the method.
>>> 
>>>
>> 
>> First of all: Thanks for the replies, I think I have enough input to
>> chew on.
>> 
>> And to reveal the purpose:
>> I'd like to match a LOT of words/strings (words with spaces) and
>> sometimes regexes against a lot of small and medium size( < 10000 byte )
>> strings.
> 
> If you are going to compare strings to strings as well as strings to
> regexes, maybe a two-tiered scheme would be better. First tier, simple
> textual comparison using a hash of strings. Second tier, regexes.
> 
> This removes the ambiguity that a particular data string might pass a string
> comparison but fail any of the provided regexes, or the opposite. It will
> also be much faster than a gigantic list of strings and regexes, all
> presented to the regex engine as though they were all regular expressions,
> even though some are simple string comparisons.
> 
> I think this (e.g speed improvement) would have been true in Perl also.
>

His regex isn't anchored to the beginning and end of the string. This makes
hashes useless for the kind of comparison he seems to want to perform.

--Ken

-- 
Ken Bloom. PhD candidate. Linguistic Cognition Laboratory.
Department of Computer Science. Illinois Institute of Technology.
http://www.iit.edu/~kbloom1/