"Simon Riggs" <simon(at)2ndquadrant(dot)com> writes:
> You are 100% right: MVs are unambiguously not allowed as part of the
> TPC-H spec - what 1.5.7 shows is how useful MVs are: they've had to ban
> them!
Ah, now I get your point. You're quite right, the TPC-H queries taken
as a repetitive workload would be a good test case for materialized
views. I misunderstood you as suggesting that we'd want to push use
of MVs into the DBT-3 benchmark as such.
regards, tom lane