fast regex

Nobody nobody at nowhere.com
Tue May 11 15:46:33 EDT 2010


On Tue, 11 May 2010 17:48:41 +1200, Lawrence D'Oliveiro wrote:

>> I was working with regex on a very large text, really large but I have
>> time constrained.
> 
> “Fast regex” is a contradiction in terms.

Not at all. A properly-written regexp engine will be limited only by
memory bandwidth, provided that the state table fits into the primary
cache.

> You use regexes when you
> want ease of definition and application, not speed.

Other way around.

> For speed, consider hand-coding your own state machine. Preferably in a
> compiled language like C.

Or use a decent regexp library.

Even if you want to use non-regular expressions (e.g. backreferences), a
decent engine will still use a DFA, bactracking only where strictly
necessary.




More information about the Python-list mailing list