David Fetter <[EMAIL PROTECTED]> writes:
> [code]
Seems pretty messy. I believe the standard way of doing this is
regression=# create function epoch_to_timestamp(float8) returns timestamptz as '
regression'# begin
regression'# return ''epoch''::timestamptz + $1 * ''1 second''::interval;
regres
On Fri, Aug 15, 2003 at 06:57:24AM -0400, Tom Lane wrote:
> David Fetter <[EMAIL PROTECTED]> writes:
> > Is there a reasonable way to add an optional param or two to mark
> > time zones, or should that just the application programmer's
> > hassle?
> The return type should be timestamptz, which mak
David Fetter <[EMAIL PROTECTED]> writes:
> Is there a reasonable way to add
> an optional param or two to mark time zones, or should that just the
> application programmer's hassle?
The return type should be timestamptz, which makes the transformation
timezone-independent.
, 2003 4:46 AM
Subject: [HACKERS] New function: epoch_to_timestamp...
> Kind people,
>
> After discussing this w/Josh yesterday, I'd like to propose a new
> function for date_time. Here's the Pl/PgSQL version, warts 'n' all.
>
> CREATE OR REPLACE FU
