Re: pg_depend explained

From: Joel Jacobson <joel(at)gluefinance(dot)com>
To: Florian Pflug <fgp(at)phlo(dot)org>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: pg_depend explained
Date: 2011-01-11 22:55:07
Message-ID: AANLkTimao2H55UZt0eiobkf8Dt3=Ng=iOnUZmfSesBEu@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

2011/1/11 Florian Pflug <fgp(at)phlo(dot)org>:
> Could you give an example of the kind of trouble you're experiencing trying
> to use a topological sort?

Let's say you have a table t and a view v.
The view v is defined as select * from t;
If we put all objects in a tree, with the public schema as the root,
both v and t will directly under the root, but in reality, v cannot be
created before t.
This is the reason why a normal topological sort doesn't work.
You have to look at the deptype and sort nodes having "internal" edges
between them differently.
The pg_dump source code of course contains all the logic necessary to
do the trick, but it's not that easy to follow.

I guess it's time for plan B, sorting based on oid, no biggie, it will
work for my purpose, but it's damn ugly.

--
Best regards,

Joel Jacobson
Glue Finance

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2011-01-11 23:06:58 pg_regress multibyte setting
Previous Message Lukas Eder 2011-01-11 22:54:19 Re: Fwd: Weird issues when reading UDT from stored function