Zac Spitzer wrote: > Bryan Love wrote: > >>Not really true. The word "terrible" is a vast overstatement. >> >>Assuming you use char (or varchar) for the UUID (char is better since the >>size is constant)... >>The difference btwn string keys and numeric keys (performance-wise) is >>pretty much null until you get up into the millions of records and even then >>it's not enough to worry about. > > > ok some simple maths, a uuid is what 30 chars or something, that means a > join using two tables with lets say 10000 and 3000 rows.. that means a > query between the two is going to be working with either 13000 bytes or > 13000 * 30 bytes just to join data... the answer there is that you are > working with 30 times the amount of data, sure there is some optimzation > but the logic goes further..
Because primary keys are always indexed, and the fields in your db used for joins should also be, it would be a join on idexes, these are usually faster then unindexed numeric joins. Indexed numeric joins are just as fast (also joining on indexes). Jesse ______________________________________________________________________ Get the mailserver that powers this list at http://www.coolfusion.com FAQ: http://www.thenetprofits.co.uk/coldfusion/faq Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/ Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists
