Tom Lane wrote:
> AFAICS, what Ryan is after would be better served by a separate tool that
> would produce a "diff" of the two tables' schemas. Even if we were
> willing to overload this error message with everything we know about the
> parent column, do you really want to fix discrepancies one column at a
> time? And what about properties that can't be uniquely associated with a
> single column, such as constraints covering multiple columns?
>
> Also, I have a feeling that a suitable tool is already out there. A
> moment's digging in the list archives found this thread with links to
> several candidates:
>
> https://www.postgresql.org/message-id/flat/561D27E7.5010906%40trustport.com
>
> and I'm pretty sure there have been other such discussions.
I remember looking some time ago, and most of all the possible
candidates were either abandoned or terribly cumbersome to use. I ran
across Euler Taveira in a conference sometime later and he told me he
had the idea of working on writing such a tool. I was pleased to see
his announcement recently:
https://www.postgresql.org/message-id/6c9d4a85-55b5-81cc-6f09-8f26a6da2072@timbira.com.br
I admit I have not looked at his code, but he had some very good ideas
for the complex cases. I think it's worth checking out what he has
done.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services