[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: indexing and pattern matching suggestions
> Absent implementations of query language in Ozone, does anyone have pointers
> or suggestions for practical ways to speed up simple pattern matching in
> large groups of similar objects?
> In my case, I'm looking at string pattern matching, the mechanics of which
> are handled just fine by regexp implementations. What I'm trying to do is
> figure out -- or borrow -- an efficient way of indexing objects by string
> value such that for any given simple regexp, you aren't scanning the whole
> group of objects, but can immediately narrow down to a subgroup.
> A half-way hack to this end would be (say) to break up your collection of
> strings into subgroups by the first three (or two or four or whatever)
> characters. However, this doesn't help you when matching something like
> "*pattern*" as opposed to "pattern*".
> So far my searching hasn't turned up any useful code...
Here is a list of stuff.
Scroll down to the gnu.regexp
I have not used but have testimony that it is good.
Hope this helps,