There are two different types of joins. The faster type is done in the database, allowing the optimizer
to choose the most efficient query plan to process multiple "fetch" requests. The slower mechanism
is to have an application generate new queries for each row, fashioning a virtual table the "hard way".
When the data is in different databases, you only have access to the latter solution, which is why most
heterogeneous joins eventually get rearchitected into what Oracle calls a "materialized view", where
the key data in one database is regularly replicated into the other for realtime use.
- V