The join queries are run with no physical hints first. Then they are run with several hints.
If a non hinted query slower than one of the hinted versions, then we know the optimizer is picking the wrong plan. The goal is to ensure that this does not happen.
We will be using the R, S, T, U data sets for this suite.