On Tue, 5 Nov 2019 14:29:00 +1300, David Rowley
wrote:
>See https://www.postgresql.org/docs/current/sql-set.html
>
>"SET only affects the value used by the current session."
>
>Also:
>
>"The effects of SET LOCAL last only till the end of the current transaction"
>
>Neither affects other sessions
Adrian Klaver writes:
> On 11/5/19 6:43 AM, Michael Shapiro wrote:
>> Can I set search_path='' (ie to a string that does not match any
>> existing schema)? Would that be the proper way to guarantee that the
>> definition for any view will always be fully-qualified?
> [ example ]
If you read th
On 11/5/19 6:43 AM, Michael Shapiro wrote:
Can I set search_path='' (ie to a string that does not match any
existing schema)? Would that be the proper way to guarantee that the
definition for any view will always be fully-qualified?
test=# show search_path;
search_pa
Can I set search_path='' (ie to a string that does not match any existing
schema)? Would that be the proper way to guarantee that the definition for
any view will always be fully-qualified?
On Sun, Nov 3, 2019 at 3:15 PM Tom Lane wrote:
> Michael Shapiro writes:
> > It seems that the definition
On Tue, 5 Nov 2019 at 11:05, George Neuner wrote:
>
> On Sun, 03 Nov 2019 16:15:23 -0500, Tom Lane
> >Why don't you just change the search path to empty before selecting?
>
>
> I'm guessing this is fine inside a transaction, but what if you're
> using a client that commits by statement? In that
On Sun, 03 Nov 2019 16:15:23 -0500, Tom Lane
wrote:
>Michael Shapiro writes:
>> It seems that the definition of a view from pg_catalog.pg_views does not
>> qualify the tables used in the view if the tables are in the current search
>> path.
>
>> Is it possible to either have the definition alway
Is that how PgAdmin does it? The views extracted by PgAdmin are fully
qualified
On Sun, Nov 3, 2019 at 3:15 PM Tom Lane wrote:
> Michael Shapiro writes:
> > It seems that the definition of a view from pg_catalog.pg_views does not
> > qualify the tables used in the view if the tables are in
Michael Shapiro writes:
> It seems that the definition of a view from pg_catalog.pg_views does not
> qualify the tables used in the view if the tables are in the current search
> path.
> Is it possible to either have the definition always qualify all tables
> independent of the search_path (or el
