On 28.10.2010 13:54, Boszormenyi Zoltan wrote:
> A little better version, no need for the heavy hash_any, hash_uint32
> on the lower 32 bits on pk_eclass is enough. The profiling runtime
> is now 0.42 seconds vs the previous 0.41 seconds for the tree version.
Actually, I wonder if we could just have a separate canon_pathkeys list
for each EquivalenceClass, instead of one big list in PlannerInfo. I'm
not too familiar with equivalence classes and all that, but the attached
patch at least passes the regressions. I haven't done any performance
testing, but I would expect this to be even faster than the hashtable or
tree implementations, and a lot simpler.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com