Fabrízio de Royes Mello <fabriziomello@gmail.com> writes:
> On Sat, Mar 1, 2014 at 2:11 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> [ re schema upgrade scenarios ]
>> Why wouldn't COR semantics answer that requirement just as well, if not
>> better?
> Just because it will replace the object content... and in some cases this
> cannot happen because it will regress the schema to an old version.
That argument seems awfully darn flimsy. On what grounds would you argue
that the script you're sourcing contains versions you want of objects that
aren't there, but not versions you want of objects that are there? If
the script is out of date, it seems more likely that you'd end up with
back-rev versions of the newly created objects, which very possibly won't
interact well with the newer objects that were already in the database.
In any case, given the existence of DO it's simple to code up
create-if-not-exists behavior with a couple lines of plpgsql; that seems
to me to be a sufficient answer for corner cases. create-or-replace is
not equivalently fakable if the system doesn't supply the functionality.
regards, tom lane