Re: Rating results after relevance. Difficult problem

send a email to google and ask them how they do? :) -- Best regards, George Nicolae IT Manager ___________________ PaginiWeb.com - Professional Web Design www.PaginiWeb.com "Andy" <news.letters@gmx.de> wrote in message news:20020508184748.15130.qmail@pb1.pair.com... > Hi there: > > I am working on a member search engine. Regarding on the criterias a user > provides (like should speek english, should be interested in arts...) the > machine querries the db and returns the results. But I would like to rate > the results. So All criterias fitt... 100%, Just 1 of 3 -> 33 % and so on. > > I did already finish my coding and works fine for a user amount of 20. But > now I generated 2000 random users and woww!!! Take about 2 minitues to > querry!! > > So there has to be a better way, right? > > What I am doing with this version, is to get all the users fitting the > criteria. Opening a temporary table store the id and the amount of fitting > criterias in it. Retriving this temporary table and sort it after the most > wanted user (100% max). > > The problem with this, is that on user takes about 0.04 s to check and store > in the temp table. If there are 2000 users fitting at least one of the > criterias thats 2000 * 0.04 s means 400 s = lots of time! Not to imagine > what happens if there are 100.000 users in the db!!! > > So the querry performance is ok, I just need to find a more effective way to > rate them. There are 3 tables. > - user table with most of the data > - language table, containing the languages the user speak (by the time of > registering they can select up to 3 languages) > - interest table, containing the interests of the users (by the time of > registering they can select up to 20 interests) > > Can anybody help on that difficult topic? I would really apreciate any hint > on that. > > Thanx in advance, > > Andy > >