> I have proposed that this should be allowed, but it is not possible
> at the moment.
i read your proposal document in the TODO.detail for inheritance. has
your proposal been accepted as an active path for postgres development?
i have a situation that calls for the scenario that Alastair outlined
above. i was about to make exactly the same post when i read his.
this thread is about postgres 7.03, but i haven't seen any evidence that
your proposals have worked their way into 7.1, as i'm using 7.1b4 and am
having exactly the same problem.
> You can refer directly to a descendant table, of course.
for me, this is not a sufficient workaround because i have a superclass
table whose subclasses hold most of the relevant data and it just
strikes me as particularly poor data design to have something like the
following setup.
foo <- sub_foo_left
<- sub_foo_right
bar <- sub_bar_left
<- sub_bar_right
where the bar subclasses exist only so they can refer directly to a
descendant table.
yes, this works, but it seems to me that the more useful way is to have
bar be able to stand alone and refer (with proper integrity) to foo
alone rather than needing it's own subclasses in order to refer directly
to the subclasses of foo.
i don't suppose there's any sort of schedule for your inheritance proposals?
-tfo