* t.dalpozzo@gmail.com (t.dalpozzo@gmail.com) wrote:
> I've a very huge table whose 1st column is a numeric value, starting
> from 0 at the 1st row and incremented by 1 each new row I inserted.
> No holes, no duplicates.
> I need to perform some very fast query based on this value, mainly
> around the last inserted rows.
> What is the best I can do? A normal index or is there a way to
> instruct the system to take advantage from that strong order?
> In theory, the number of the row is the only info the system would
> need to directly access that row.
> So I'd like to avoid useless complex indexes if possible.
A BRIN index should work pretty well in that scenario.
A btree index would most likely be better/faster for query time, but
more expensive to maintain.
Thanks!
Stephen