On Mon, Aug 3, 2020 at 8:59 AM Konstantin Knizhnik <k.knizh...@postgrespro.ru> wrote: > May be generated and identity columns are good things. I have nothing > against them. > But what preventing us from providing the similar behavior for > serial/bigseries types?
Backward compatibility seems like one good argument. It kind of sucks that we end up with cases where new notions are introduced to patch up the inadequacies of earlier ideas, but it's also inevitable. If, after 25+ years of development, we didn't have cases where somebody had come up with a new plan that was better than the older plan, that would be pretty scary. We have to remember, though, that there's a giant user community around PostgreSQL at this point, and changing things like this can inconvenience large numbers of those users. Sometimes that's worth it, but I find it pretty dubious in a case like this. There's every possibility that there are people out there who rely on the current behavior, and whose stuff would break if it were changed. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
