| CBO vs RBO Explain plan [message #65972] |
Fri, 11 February 2005 00:39 |
Kilburn High Rd
Messages: 1
Registered: February 2005
|
Junior Member |
|
|
Folks,
I know this is an old chestnut .... but here we go ..
I have one piece of SQL. If I run this with a /*+ RULE */ hint then it chooses an explain plan that is different to when I run it without the /*+ RULE */ hint. - Nothing too surprising in that, but what I am not sure about is why the explain plan produced with the /*+ RULE */ hint should be orders of magnitude faster than without the /*+ RULE */ hint.
I seem to remember being at various conference sessions where the black art of CBO is discussed. What I really want to know is ... Has anyone written a white paper or "How to" on solving these sort of problems. i.e. a sort of checklist or system process of what you should check and when, in the order of likely culpability?
I know that with these problems, one starts by checking stats on tables, then looking at differences in the explain plans, but I am sure there are many other things that one would look for after these.
So ... in summary, Does anyone know of a document that provides a framework for working through unexpected CBO vs RBO performance differences?
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
