[rules-users] Queries vs Rules with same LHS

I'm looking for more information on how queries are handled/processed in the Drools rule engine v. 4.0.7? I am specifically interested on a computational cost of a query compared to a rule with the same LHS. As I understand a rule will match on any single fact that fulfills LHS. On the other hand a query will provide all facts that fulfill its LHS. How much more does this cost? Is there a best practice doc or article on this type of information? Dave