RULE vs. COST OPTIMIZATION & STATIS

We are currently having a debate in our shop on cost based vs. rule based optimization and the collection of statistics on our tables and indexes. The status quo has been rule based optimization (which is out default). Collection of statistics has not been enabled. We are however starting to use hints in our SQL which at times has improved performance.

It was my understanding that the use of hints invoked cost based optimization - which seemed to benefit from the collection of statistics.

I am being told that a single SQL statement can use a mixture of cost and rule based optimization, i.e. joins.

I have 3 questions:
1. Does the collection of statistics impact rule based optimization ?
2. Does the collection of statistics cause an outage to the tables involved; couldnt concurrent read happen ? 3. Is it true that a single SQL statement can use both rule and cost based optimization ??

Please email me at Jeff_Kluth_at_metcapw1.ccmail.compuserve.com. ahev My question is: If you use hints, wouldnt you be better off collecting statnshouldnt you cant you Received on Sun Aug 06 1995 - 00:00:00 CEST