On 27 Jul 2009, at 10:44pm, CityDev wrote: > Over the intervening years I can't ever remember denormalising data > (even > when dealing with eg 13 million insurance customers in a table). Is > it OK > nowadays to say always aim to be fully normalised - modern RDBMSs are > usually powerful enough to cope with most anything?
Performance/space/time/money payoff calculation. The smallest mundane laptop these days comes with a 120 Gig hard disk. No real need to economise on file size, so if you need fast processing you might use lots of space. On the other hand if you need to cart your entire dataset around on a USB Flash drive you might want to save space. The advantage is that most programmers don't have to care which way they go: they write the program any way it occurs to them. The disadvantage of this is that when they hit the first job where the difference does matter, they don't know what to do about it. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
