On Sat, 2003-08-30 at 10:47, Rob Nagler wrote:
> Tom Lane writes:
[snip]
> enough. When I add this index, I will slow down inserts (about
> 20K/day) and increase data size (this is the second largest table in
[snip]
Since I gather that this is a web-site, can we presume that they
are clumped into an 8 hour range? 20,000/8 = 2,500/hour, which
is 41.67/minute. If you can't do .69 inserts/second, something
is wrong, and it ain't hardware, and it ain't Postgresql...
> > PS: does server_id really need to be NUMERIC? Why not integer, or at
> > worst bigint?
>
> It is a NUMERIC(18). It could be a bigint. What would be the change
> in performance of this query if we changed it to bigint?
http://www.postgresql.org/docs/7.3/static/datatype.html#DATATYPE-INT
http://www.postgresql.org/docs/7.3/static/datatype.html#DATATYPE-NUMERIC-DECIMAL
Scalars are faster than arbitrary precision types. Small (32 bit)
scalars are faster than bit (64 bit) scalars on x86 h/w.
--
-----------------------------------------------------------------
Ron Johnson, Jr. ron.l.johnson@cox.net
Jefferson, LA USA
"Adventure is a sign of incompetence"
Stephanson, great polar explorer
