On Fri, 2008-07-18 at 11:07 +0100, Gregory Stark wrote:
> "Simon Riggs" <simon@2ndquadrant.com> writes:
> hash lookups can in theory be O(1).
I'm not sure whether that applies here? I'm interested in how *this*
patch will work, not in more generic algorithm theory.
To patch authors: Can we please see a table showing expected number of
logical I/Os (i,e, block accesses) for btrees and hash indexes
e.g. for 100-byte rows...
rows btree hash
---- ----- ----
10^2
10^3
10^4
10^5
10^6
10^7
10^8
-- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support