What is "data-intensive business logic" in the context of Java stored procedures?

Question

I'm reading an Oracle book, which states:

Java stored procedures are, of course, written using Java, and they facilitate the implementation of data-intensive business logic using Java.

What is "data-intensive business logic"? Would OLAP be an example and if not what would be?

Answer 1

DB people love to say that all your "business logic"* should be in the database. In practice, I haven't seen this work, but even if you don't subscribe to this principle there can be cases where you need to do intensive data manipulation; say, transforming text-based records in such a way that requires data from multiple tables per document transformation.

Doing these kinds of things in the database (i.e. with stored procedures) saves you the serialization-network-deserialization-possible-ORM-etc hit, so it can occasionally be sensible. If you've already decided to do this, then doing it in Java if you have the option is 10000x better than doing it in PL/SQL which invokes cthulhu.

*business logic meaning e.g. expiring/renewing customer subscriptions. data-intensive business logic might be some kind of data warehousing/statistics, I suppose.