RE: SQL for performance (overnight batch jobs)

An important function of batch jobs (large and/or overnight) is
providing batch control totals and other tools accountants (the fact
that "tools" and "accountants" are contiguous in this sentence should
not be construed as a political statement) rely on to demonstrate system
integrity.  In addition, it's a useful form of work management (unless
you like printing 35,000 invoices interactively).

The revenue accounting environment in the transportation/logistics
environment is complex: there are pricing, discounting, revenue split,
revenue allocation, and fuel price fluctuations driving every invoice in
a high-volume (100/day-20,000/day) environment. 

Real-time processing is the solution for just about everything but
accounting.  Cutoff dates (period end, etc.) are an integral part of the
"measure and compare" nature of accountancy.

I don't understand how the use of SQL could improve performance in an
otherwise well-designed (proper use of LF's, etc.) batch application.
SQL may offer some flexibility with complex joins (for reporting, of
course) but that's about all I can think of. 

-reeve