Martijn van Oosterhout <kleptog@svana.org> writes: > I've considered whether it's worthwhile going to other way: getting the > IndexScan executer node to uses getmulti to reduce index AM overhead. > But that requires backward scan support also...
I think Heikki got most of the low-hanging fruit already with that patch for page-at-a-time scanning in btree. There's some wasted overhead just from multiple levels of function call, but I doubt it's really a big deal anymore. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 3: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faq