Kirk wrote: > The argument is to name it "id" or another name like "client_id". > I like client_id because then the primary key name matches the > foreign key in another table. > My coworker likes just the name id because then he can have a base > class called getId()
Personally, I always have a column in each table called "id", and then if I'm using it as a foreign key in another table, I reference it as, e.g., "client_id". By that I mean I might have a client table: CLIENT id integer primary key auto_increment blah blah other data fields and then a table that contains, say, documents for that client: CLIENT_DOCUMENT id integer primary key auto_increment client_id integer default 0 blah blah other data Here, there might be multiple client documents, and each of them will contain a client_id value matching the id field of a client record. $0.02, Jeff
pgpFEQkR6cUvF.pgp
Description: PGP signature
/* PLUG: http://plug.org, #utah on irc.freenode.net Unsubscribe: http://plug.org/mailman/options/plug Don't fear the penguin. */
