On Sat, 12 Sep 2009, Tom Lane wrote:
> Everybody in the world is going to want their own little problem to be
> handled in the fast path. And soon it won't be so fast anymore. I
> think it is perfectly reasonable to insist that the fast path is only
> for "clean" data import.
The extra overhead is that when you hit the checks that are already in the
code, where the row would normally be rejected, there's a second check as
to whether that particular problem is considered OK or not. There won't
be any additional overhead for clean imports. As I was pointing out in
one of the messages in this thread, all of the expensive things you need
are already being done.
As for "everybody in the world" wanting a specific fix for their private
problems, I assure that everything I suggested comes up constantly on
every legacy data conversion or import job I see. This is not stuff that
fits a one-off need, these are things that make it harder for people to
adopt PostgreSQL all the time. I wouldn't care about this one bit if
these particular issues didn't ruin my day constantly. Consider the fact
that we're looking at three samples of people who have either already
written a patch in this area or considered writing one showing up just
among people on the hackers list. That should be hint as to how common
these requests are.
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD
