RE: Optimizer theory: Question on access paths for outer joins

2001-04-11 Thread Steve Adams
Hi Jay, Under rule based optimization deficient (outer joined) relations are always last in the join order. In general cost based optimization will do the same, but it is not bound to do so. For example, if T2 below is a complex view, then it cannot be merged into the parent query because it is

RE: Optimizer theory: Question on access paths for outer joins

2001-04-11 Thread Steve Adams
Hi Jay, Under rule based optimization deficient (outer joined) relations are always last in the join order. In general cost based optimization will do the same, but it is not bound to do so. For example, if T2 below is a complex view, then it cannot be merged into the parent query because it is

Re: Optimizer theory: Question on access paths for outer joins

2001-04-10 Thread Regina Harter
I am a little confused. Is not PHONE, in this query, the one WITHOUT the +? At 11:25 AM 4/10/01 -0800, you wrote: Hi all, I was just trying to figure out why in an outer join Oracle prefers to access the table with the (+) first. I would have thought that the table from which all the data was

RE: Optimizer theory: Question on access paths for outer joins

2001-04-10 Thread Hillman, Alex
if ac.id column is indexed for outer join to use this index this column should be not ull or condition should be added ac.id is not null or ac.id0 etc. It is because an index will not have rowid's for rows when this column is null. Alex Hillman -Original Message- Sent: Tuesday, April

RE: Optimizer theory: Question on access paths for outer joins

2001-04-10 Thread Miller, Jay
No, I was the one who was confused when I retyped the query to send it. It should read: select stuff from inlineview1 T0, inlineview2 T1, inlineview3 T2, account ac, phone ph where T0.generic_id = T1.account_id and T1.account_id = T2.account_id (+)