"John D. Burger" <john@mitre.org> writes:
>> mycol timestamp default now()
>> or
>> mycol timestamp default current_timestamp
> And both of these return start time of the current transaction, yes?
> Is it the case that there is no SQL-standard way to get the current
> time?
AFAIK the spec doesn't really define the meaning of current_timestamp
all that closely. I believe there's a patch in the queue to add some
additional nonstandard functions for current time and start of current
interactive command, though.
regards, tom lane