I think, it should be new node in executor. Planner select classic index scan or new functional index scan. 2011/9/25, Tom Lane <tgl@xxxxxxxxxxxxx>: > =?ISO-8859-2?Q?pasman_pasma=F1ski?= <pasman.p@xxxxxxxxx> writes: >> My english is not perfect, by accumulative i think about monotonically >> increasing function. > > Oh, I see how that would work. I can't get real excited about it > though. The use-case seems a bit narrow, and the amount of complexity > added to the btree search mechanism (thereby slowing down all btree > searches) would be significant. Furthermore, unless f() is pretty cheap > to evaluate, you'd end up preferring an index on f(x) anyway, because > that can be searched without any new evaluations of f(). > > regards, tom lane > -- ------------ pasman -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
