Chris Bitmead <chrisb@nimrod.itg.telstra.com.au> writes:
> Tom Lane wrote:
>>>> 1) An imaginary field in every tuple that tells you the class it came
>> This is a good idea, but it seems to me that it'd fit into the system
>> traditions better if the pseudo-field gave the OID of the source
>> relation.
> What do you think about having both? I know you can go from one to the
> other by joining with pg_class, but that's too inconvenient, and I can't
> make up my mind which is the better "system tradition" either.
If we can implement it as I sketched before, there's no reason not to
offer both, since either one would create zero overhead for any query
not using the feature.
I'll comment on the other issues later ... but I will say that I don't
think it's acceptable to add *any* overhead to standard-SQL queries
in order to support inheritance better. The vast majority of our users
want SQL performance and don't give a damn about inheritance. We have
to pay attention to that.
regards, tom lane
