I/O Parallelism

[login to unmask email]

I/O Parallelism
Is it possible to get I/O parallelism on a LOAD RESUME YES job that loads a
partitioned tablespace?
If CPU usage is over 70%, does the benefit of CPU parallelism get
diminished?

We currently load our data warehouse by date and thus we partition by date.
Would there be any advantage
to partitioning our tablespaces by using a random key (hash) to enable the
partitions to be loaded equally? Does this
help out the load job or just enable queries to complete in a faster manner
or both?