Re: pg_dump of table including check rule fails to restore - Mailing list pgsql-bugs

From Lewis Foti
Subject Re: pg_dump of table including check rule fails to restore
Date
Msg-id MJEPJJCNLEBMDJFMDDABCEHOCDAA.lewis.foti@mentation.com
Whole thread Raw
In response to pg_dump of table including check rule fails to restore  ("Lewis Foti" <lewis.foti@mentation.com>)
Responses Re: pg_dump of table including check rule fails to restore
List pgsql-bugs
Further to my original message I realise that the problem was due to the
text of the rule still referring to columns, src_section_id and
dst_section_id by their original names, parent_section_id and
child_section_id respectively. So the nature of the bug changes to that
pg_dump does not correctly export CHECK rules where the name of the
column(s) referred to has changed. Now IMHO that is somewhat subtle.

regards

Lewis

-----Original Message-----
From: Lewis Foti [mailto:lewis.foti@mentation.com]
Sent: 16 December 2003 18:26
To: pgsql-bugs@postgresql.org
Subject: pg_dump of table including check rule fails to restore


Hi

I have a schema which includes one table with a CHECK constraint. After the
schema is dumped (with pg_dump) as ascii text attempts to recreate it using
the psql -f <filename> command cause this one table, called navigaion, to
fail to create. This is using postgresql 7.3.3 on Redhat 9.0.  The fragment
of the dump is below. Is there a cure for this problem?

regards

Lewis

--
-- TOC entry 6 (OID 154551)
-- Name: navigation; Type: TABLE; Schema: public; Owner: wallet
--

CREATE TABLE navigation (
    src_section_id integer NOT NULL,
    dst_section_id integer NOT NULL,
    rank smallint NOT NULL,
    CHECK ((parent_section_id <> child_section_id))
) WITHOUT OIDS;

Lewis Foti

e: lewis.foti@mentation.com
m: +44 (0)7771 535943
w: www.mentation.com

pgsql-bugs by date:

Previous
From: "Lewis Foti"
Date:
Subject: pg_dump of table including check rule fails to restore
Next
From: Lauris Bukšis
Date:
Subject: unicode upper and lower functions