GilaMonster (5/19/2012)
You are highly unlikely to achieve anything other than adding complexity and wasting time. Partitioning is not a performance optimisation (mostly), it's for ease of maintenance, loading and deleting rows.Partitioning a table so that is it useful requires a careful choice of partitioning column based on the data and the queries, and a semi-random hash is highly unlikely to be useful.
Honestly, don't go this route.
What about for tables with no natural partitioning key and a low cardinality on all candidates?
There are no special teachers of virtue, because virtue is taught by the whole community.
--Plato