Think about Loose Coupling | |
PerlMonks |
comment on |
( [id://3333]=superdoc: print w/replies, xml ) | Need Help?? |
that algo limits the scope of regex matches to individual terms. Allowing for post-wildcarded keywords stem* is relatively easy, especially if you limit the prefix to a minimum of (say) 3-chars or more. Where I had to provide for a pre & post wildcards, I built a trigram index (and insisted upon a minimum of 3 contiguous chars). That worked well, but required a substantial sized index. The one search engine I am familiar with that does allow full regex including term spanning is Google's codesearch. I'd love to see insights into how they do that. It's quite interesting to see the difference in the time it takes to do
Really quite remarkable even given the massive parallelism. They must do some pretty clever parsing of the regexes in order to avoid searching the entire dictionary. I guess the code corpus is considerably smaller than the main set which will help a lot. That said: It makes you wonder why the OP doesn't just use Googles search engine? Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
In reply to Re^3: What DB style to use with search engine
by BrowserUk
|
|