Re: database connections

On Mon, July 24, 2006 1:33 pm, Ryan A wrote: > Hey Rob, > Thanks for replying. > > >> It's usually a sign of poor programming and/or >> purist OOP programming. >> >> When I say purist OOP programming... >> I saw >> one really retarded >> implementation of this kind of system where an >> excess of 20000 queries >> were issued to the database -- on a homepage >> nonetheless :/ > > > That IS retarded, I wonder why someone would want to > do that. Because they learned from Java which has an entirely different architecture with a shared cached DB store across the application servers, wherein such a technique makes a whole hell of a lot more sense. And because they don't REALLY understand the power of OOP, and are too literal-minded when architecting class inheritence -- But that describes about 99% of the so-called OOP developers "out there" so there ya go. > I was curious about this because I am working on a > project (with other team players) and we have a way of > building something with either lots more (complicated) > code and fewer database calls or less code and > multiple tables. > > If we take the second option (multiple tables) I am > talking about maybe 15 database calls per page, and > the site will get around (i guess) 300-750 requests > for a page a minute at is peak. Measure it and see. I've seen times when 15 DB calls was a hell of a lot cheaper than 1 DB call, depending on the indexes in the DB and the query and... There's just no way to predict this without a detailed analysis of the DB schema and the queries and the hardware and... It's almost-for-sure faster for you to write 2 prototype examples with realistic-sized data and throw Apache benchmark at them to find out your answer. > I wouldnt be going to those extremes, was thinking of > around 5-15 queries per page. 15 queries per page, with a good schema and reasonable queries is chump-change. 15 bad queries on a page will kill you before you even launch. :-) -- Like Music? http://l-i-e.com/artists.htm