On Monday 21 December 2009 5:17:49 pm Scott Marlowe wrote:
> On Mon, Dec 21, 2009 at 6:14 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> > Scott Marlowe <scott.marlowe@gmail.com> writes:
> >> On Mon, Dec 21, 2009 at 6:10 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> >>> Well, yes, because only the first one is visible. The second one is
> >>> masked by the first.
> >>
> >> But the docs say that ALL objects in the schema path will be shown.
> >> So, my point stands, either the docs are wrong, or the behaviour is.
> >> I'd think it's the docs.
> >
> > It says the *visible* objects will be shown. Ones that are masked
> > aren't any more visible than if they were in some other schema
> > altogether: either way, if you want to reference such an object in
> > a SQL statement, you'd have to schema-qualify it.
>
> Ahh, right, it's about visibility. Hadn't caught that part.
I think that is where the biggest misunderstanding lies. The problem is that
people new to the database may not fully understand what visible means in this
context. I know this tripped me up the first time I encountered the identical
name situation. I made the same assumption the OP did, the tables where in the
search_path and I had permissions on them so they should be 'visible'. It took
some digging around to find the correct answer. I not sure how the best way is
to clarify that in the psql documentation.
--
Adrian Klaver
aklaver@comcast.net
