Maciek Sakrejda wrote:
>> In general, no. That's because resultset columns can be "entangled" in more complex way than "column a of table b".
>
> Actually, this *is* defined at the protocol level:
> No idea whether JDBC exposes this or whether the server actually sends
> the correct oids, or if this is just part of the protocol as a
> future-proofing attempt.
The driver exposes this via PGResultSetMetadata.getBaseTableName(), and
I believe the server does send useful data here.
-O