Re: [HACKERS] Inherited constraints and search paths (was Re: - Mailing list pgsql-general

From Tom Lane
Subject Re: [HACKERS] Inherited constraints and search paths (was Re:
Date
Msg-id 3776.1116624678@sss.pgh.pa.us
Whole thread Raw
In response to Re: [HACKERS] Inherited constraints and search paths (was  (Berend Tober <btober@seaworthysys.com>)
List pgsql-general
Berend Tober <btober@seaworthysys.com> writes:
> However, when I do a select from a view, which itself does a select from
> a parent table, the query result does include the child table rows,
> i.e., the SQL_INHERITANCE  setting is ignored in this situation. Should
> the SQL_INHERITANCE  setting still rule?

I believe what matters is the sql_inheritance setting that was in force
when the view was created.  You should be able to replace the view and
get it to do what you want.

(Hmm ... which suggests that we have still another pg_dump issue,
because views will be dumped using ONLY or no marker, and so reloading
them into a server with SQL_INHERITANCE off would result in a change in
the view behavior.  Probably we need to put "SET SQL_INHERITANCE = ON"
into the standard prologue of pg_dump scripts.)

            regards, tom lane

pgsql-general by date:

Previous
From: Matthew Hixson
Date:
Subject: GCC 4.0 on Mac OS X
Next
From: Tom Lane
Date:
Subject: Re: GCC 4.0 on Mac OS X