DB design document

[David Merrill]

On Wed, Dec 13, 2000 at 10:02:35PM -0500, Roland Roberts wrote:
> >>>>> "dm" == David Merrill <dmerrill@lupercalia.net> writes:
> 
>     dm> On Wed, Dec 13, 2000 at 06:41:34PM -0800, Dave Peticolas wrote:
>     >> David Merrill writes:
>     >> > On Thu, Dec 14, 2000 at 12:08:24PM +1000, Phillip Shelton wrote:
>     >> > > 
>     [...]
>     >> > > Do we need to worry about transaction_time here?
>     >> > 
>     >> > I'm not sure about Postgres, but in Oracle a date field is really a
>     >> > date+time field. But I'll change the doc so it's obvious.
>     >> 
>     >> GnuCash currently allows any date/time field to be set at nanosecond
>     >> resolution. It seems a bit overkill to me, so we might want to think
>     >> about dropping that, but it's something to keep in mind because right
>     >> now the api for it is there.
> 
>     dm> We will be limited by the precision of the database. I'm not sure
>     dm> about Postgres' precision, but it is probably millisecond or so.
> 
> Oracle date type is date+time and is accurate to one second, at least
> on the hosts I've had access to.  Postgres has date, time, and
> timestamp types.  "date" is for dates only, accurate to the day;
> "time" is for time of day only, accurate to the microsecond, and
> timestamp is for date+time, accurate to the microsecond (range from
> 4713 BC to AD 1465001).

Thanks. I'll be using timestamp then.

-- 
Dr. David C. Merrill                     http://www.lupercalia.net
Linux Documentation Project                dmerrill@lupercalia.net
Collection Editor & Coordinator            http://www.linuxdoc.org
                                       Finger me for my public key

Humor in the Court:
Q: Now, you have investigated other murders, have you not, where there was
   a victim?