Re: Proposal for SYNONYMS

From: Ragnar <gnari(at)hive(dot)is>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Proposal for SYNONYMS
Date: 2006-03-10 23:21:35
Message-ID: 1142032895.18656.194.camel@localhost.localdomain
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On fös, 2006-03-10 at 16:51 -0500, Tom Lane wrote:
> Kris Jurka <books(at)ejurka(dot)com> writes:
> > One key difference would be that synonyms track schema updates, like
> > adding a column, to the referenced object that a view would not.
>
> That raises a fairly interesting point, actually. What would you expect
> to happen here:
>
> CREATE TABLE foo ...;
> CREATE SYNONYM bar FOR foo;
> CREATE VIEW v AS SELECT * FROM bar;
> DROP SYNONYM bar;
>
> With the implementations being proposed, v would effectively be stored
> as "SELECT * FROM foo" and thus would be unaffected by the DROP SYNONYM.
> Is that what people will expect? Is it what happens in Oracle?

At least on Oracle8, you could create a synonym on a
non-existing table, so if table FOO does not exist:

CREATE SYNONYM BAR FOR FOO; -- no error
SELECT * FROM BAR; -- error "synonym translation is no longuer valid"
CREATE TABLE FOO (a varchar2(10));
INSERT INTO FOO VALUES ('a');
SELECT * FROM BAR; -- no error
CREATE VIEW X AS SELECT * FROM BAR;
SELECT * FROM X; -- no error
DROP SYNONYM X; -- no error
SELECT * FROM BAR; -- error

gnari

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Jaime Casanova 2006-03-11 06:23:35 There is a problem with the download site?
Previous Message Tom Lane 2006-03-10 23:20:59 Re: problem with large maintenance_work_mem settings and