Important update to Project Part 4, Part A (Indexes) on Assignment #5:

It turns out students are having some difficulty finding realistic
AuctionBase queries that run significantly faster with indexes than
without.  There are two possible reasons:

(1) Your tables may not be large enough for indexes to truly make a
    difference.

(2) Enabling the "better" Oracle optimizer may be causing Oracle to do
    much better than we expected on queries without indexes, offsetting
    the speedup obtained by using indexes properly.

Suggested strategy:

Forget about realism, forget about lots of indexes, and maybe even
forget about your original AuctionBase tables.  Here are a couple
strategies that should work:

(a) Try a join of several tables, or join a table to itself several
    times, with a really big result.  The join conditions might be
    equalities or inequalities on attributes with a simple type such as
    integer or float (a price, for example).  Try it with and without
    indexes on the join attributes.

(b) Create a temporary table for the purposes of experimenting with
    indexes.  Populate it by using joins or cross-products on your
    AuctionBase tables.  Make it really big, much bigger than any existing
    tables.  Try a selection query on it, with and without an index on the
    relevant attribute(s).

Although we'll be more impressed with strategy (a) than (b), needless
to say we'll give full credit for using either of these strategies
even though they deviate from the original assignment.
(Another bump in the road, please accept our apologies.)