Re: pg_dump: Sorted output, referential integrity

At 10:34 11/12/01 -0500, Jan Wieck wrote:
>
> We don't want to define the constraints with ALTER TABLE
> because this means checking data on restore that doesn't need
> to be checked at all (in theory). If he has a crash of a
> critical system and restores from a dump, I bet the farm that
> he wants it FAST.

This is just an argument for (a) using ALTER TABLE (since it will
also prevent PK indexes being created, and make it FASTer), and
(b) the ability to 'SET ALL CONSTRAINTS OFF' (or similar) to
prevent the ALTER TABLE from forcing validation of the constraint.

The current situation of creating constraint triggers is IMO not
acceptable in the long term.

There are also enough people who just restore one table to warrant
the ability for pg_dump to optionally run with constraints ON.

----------------------------------------------------------------
Philip Warner | __---_____
Albatross Consulting Pty. Ltd. |----/ - \
(A.B.N. 75 008 659 498) | /(@) ______---_
Tel: (+61) 0500 83 82 81 | _________ \
Fax: (+61) 0500 83 82 82 | ___________ |
Http://www.rhyme.com.au | / \|
| --________--
PGP key available upon request, | /
and from pgp5.ai.mit.edu:11371 |/