Fast partial-word matching? Does it exist?

I'm looking for a database (or database plug-in) that can do arbitrary substring searches very quickly on a mostly-text database, or possibly even regexp searches. Whole-word indexing is not enough, here; I have to support partial-word matching at the very least. Maybe I'm just not looking hard enough, but it seems most databases I've surveyed will just read through all records sequentially! Not very useful when you have a multi-gigabyte database. I need something that will build some sort of index (e.g. a suffix tree, or whatever) to support very fast arbitrary substring searches. I'd prefer something that runs under linux, but I'd also be willing to consider other platforms if nothing suitable is available under linux. And if a pre-canned solution isn't available, I'd also take a look at any C, C++, or Perl libraries.

Thanks for any help,

Damon Hastings Received on Wed Dec 08 1999 - 21:27:14 CST