I may be missing something here, but haven't we always stated that
using 'SELECT *' should be frown'd upon for the most part? Is there a
reason why adding a column/field to an existing view should be considered
a bad thing?
As long as we don't remove existing colums that an app could be using, but
only adding a column, there shouldn't be any issues with backwards
compatibility, shoudl there?
On Sun, 23 Jan 2005, Josh Berkus wrote:
> Tom,
>
>> Any new schemas introduced by PG itself will be named pg_something.
>> This is not open to negotiation --- it's what we've promised to users
>> to avoid tromping on their schema namespace.
>
> I can see the sense in that. So, there's four ways I can see to do things:
>
> 1) leave the existing views (pg_tables, pg_views, etc.) the way they are
> except for adding columns. Create new views based on the naming scheme of
> the old.
>
> 2) create new views in pg_catalog, using new names. The problem with this is
> that the most intuitive names (pg_tables, pg_views) are taken by the old
> views and I'm not sure what to name the new ones.
>
> 3) create a new schema with the system views in it, called for example
> pg_system_views. This seems cluttered to me; a whole new schema just for a
> dozen views?
>
> 4) ignore backwards compatibility and just re-write the old views. I can
> hear the shouting already ...
>
> So, a choice of annoying options. Does anyone else on the channel have
> opinions?
>
> --
> Josh Berkus
> Aglio Database Solutions
> San Francisco
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>
----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664