announce: the "Join Over Histograms" paper

Dear all,

the paper below presents the results of an investigation about
"the formula used by the Cost Based Optimizer to estimate
the cardinality of an equijoin, when both the columns referenced in the join predicate have histograms collected."

Some practical scenarios (and possible pitfalls) are illustrated as well.

http://www.adellera.it/investigations/join_over_histograms/index.html

I hope that this paper turns out to be useful and of interest.

regards
Alberto Dell'Era

-- 
Alberto Dell'Era


"dulce bellum inexpertis"

--
http://www.freelists.org/webpage/oracle-l