Re: fixes for date_part micro/millisecond precision

On 24 Nov 2001 at 15:32 (-0500), Tom Lane wrote:
| Brent Verner <brent(at)rcfile(dot)org> writes:
| > ... if I insert a timestamp of
| > '2001-1-1 11:11:11.12341234-05' I /really/ want to get back '12341234'
| > when I ask for the microseconds that I stored,
|
| You cannot expect to get that back exactly, because *the precision is
| not there* (and yes, the term here is precision not accuracy). Right
| now (late 2001) we have about seven digits of precision to the right
| of the decimal point in a timestamp. As we get further away from the
| 2000-01-01 origin, precision will drop; it'll be six digits or less
| by 2010. The further you go from 2000 in either direction, the worse
| the precision.

ah, I see. I wasn't thinking about the limited range that could be
accurately stored in the double... :-(

| This is an inherent property of float arithmetic and can't be papered
| over with display hacks, at least not without losing more than you gain.
|
| > Is there a better place to 'force' the accuracy of the fractional
| > second part that would be less prone to introduce other problems?
|
| You cannot force accuracy that isn't there.

thanks, and sorry for the hassle.

wandering-off-to-try-making-Timestamp-a-struct-ly yours,
brent

--
"Develop your talent, man, and leave the world something. Records are
really gifts from people. To think that an artist would love you enough
to share his music with anyone is a beautiful thing." -- Duane Allman