Thread: BUG #6047: prepare p1 as select 'now'::timestamp; then "execute p1" many times, they return the same time

The following bug has been logged online:

Bug reference:      6047
Logged by:
Email address:      wcting163@163.com
PostgreSQL version: 9.0.4
Operating system:   WinXP 32bit
Description:        prepare p1 as select 'now'::timestamp; then "execute p1"
many times, they return the same time
Details:

when i execute the following statements:

prepare p1 as select 'now'::timestamp;
execute p1;
execute p1;

every time i execute statement "execute p1", they all return the same time,
the time when "prepare p1 ..." is created.
but i expect it return the changing time ---- the current time.

1). although i kown that, "the system will convert **now** to a timestamp as
soon as the constant is parsed",
i think this is a bug.

2). maybe you suggest me:
prepare p1 as select now()::timestamp;

then "execute p1" will return the current time.
this is because now() is a builtin function, so we can change "select
'now'::timestamp" to "select now()::timestamp";

but, what about "prepare p1 as select 'today'::timestamp"??
today() is not a builtin function, we can't change it to "select
today()::timestamp";


So, the implementation method of "now" and "today" is different, and
prepare p1 as select 'today'::timestamp;
execute p1;-- will always return the day when the "prepare p1 ..." is
created, even we're in another day.

On 02/06/11 10:53, wcting163@163.com wrote:
> but, what about "prepare p1 as select 'today'::timestamp"??
> today() is not a builtin function, we can't change it to "select
> today()::timestamp";

The keyword-to-timestamp conversions are ugly historical hacks.

Use the SQL-standard current_date, current_time, and current_timestamp
keywords instead.

http://www.postgresql.org/docs/current/static/functions-datetime.html#FUNCTIONS-DATETIME-CURRENT

--
Craig Ringer

"" <wcting163@163.com> writes:
> 1). although i kown that, "the system will convert **now** to a timestamp as
> soon as the constant is parsed",
> i think this is a bug.

Sorry, it's not a bug, and we're not going to change it.

> but, what about "prepare p1 as select 'today'::timestamp"??
> today() is not a builtin function, we can't change it to "select
> today()::timestamp";

You can get that by using date_trunc on the result of now(); or there's
current_date, which is actually SQL-standard unlike these other things.

            regards, tom lane