> I'm not sure about the algorithms you're using, but it sounds OK to me. > Are > you experiencing a significant performance hit using your algorithm?
I'm not seeing any performance hit at the moment. It seems to be very fast in fact. I could have implemented the levenshtein algorithm all by itself but this would have then had to go through every link stored in the DB whereas, using DIFFERENCE, it pulls out a tiny subset (3 or 4 rows at most) so the levenshtein algorithm has much less work to do. Also, if the DIFFERENCE only returns one hit, the levenshtein algorithm is skipped entirely. > Why are you looking for a better way? It's my first stab, just one solution to a problem where we are seeing where people are transposing and/or missing/adding characters etc when they are typing in a URL from printed documentation. Being a small shop, I sometimes like to get opinions on other ways of doing things... > An alternative solution, I believe, would be to use Lucene, and build > an > index of your SES url's and then I believe you can use "Did you mean" > (similarity scores) functionality to grab the most likely URL. This > might > perform better then the SQL equivalent. I hadn't considered this it may well be an option. I'll take a look thanks... :) Paul ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| ColdFusion MX7 by Adobe® Dyncamically transform webcontent into Adobe PDF with new ColdFusion MX7. Free Trial. http://www.adobe.com/products/coldfusion?sdid=RVJV Archive: http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:276037 Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
