On Mon, Nov 02, 2009 at 08:10:47PM -0800, Greg Stark wrote:
> As far as i know all of these actually work with doubles
> though, so you'll lose precision.
IEEE 754 floating point numbers (i.e. float8 or "double precision" in
PG) are defined to have a 52 bit significand and hence can store integer
values up to 2^52 without loss of precision. How good PG is numerically
I'm not sure, but you should be able to get pretty close to this range.
--
Sam http://samason.me.uk/