News for hackers. Stuff that matters.
Re:Understanding Postgres Performance by Someone on Wednesday October 03, 2012 @02:29AM (#4606122)
Thanks. But a disk seek is not a small constant. I would almost describe that as "in theory log(n), in practice a disk seek, so half the speed" (assuming that the record can be written with one disk seek)
Re:Understanding Postgres Performance by jeltz on Friday October 05, 2012 @07:09AM (#4616387)
A hash index will also require one disk seek which makes b-tree indexes and hash indexes have more similar performance than the theoretical O(log n) vs O(1) would indicate.