Hi Elein,
elein wrote:
> The problem is that after a dump and reload of
> a table hierarchy there are different values in
> pg_attribute.attislocal.
>
> A quick grep shows few references to attislocal.
> But I cannot say for sure it is unused since it is
> documented. However, I'm looking at a db diff
> tool and there it does matter.
It's not unused, though it's not a hot spot (it's only used to prevent
you from dropping an inherited column). I can't reproduce your problem
here though -- the dump comes out just like the tables I produced. Care
to provide the original SQL script?
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.