EXPERT RESPONSE
SQL statement tuning and the differences between RBO and CBO is quite a
topic. Hopefully, the following will get you started. But more study
will be necessary as this is only a cursory glance at the subject.
Whenever Oracle receives a SQL statement, it must decide the best
manner in which to retrieve the data from the database. It is the job
of the Optimizer to determine the most efficient method to use. There
could be many different access paths to the data. In the earliest
versions of Oracle, the Optimizer used a set of rules (RBO) to decide
how to execute a given SQL statement. One rule stated than an index
would be used in favor of a full table scan. So let's look at an
example table of employees:
EMP
--------------------
ID NUMBER PRIMARY KEY
NAME VARCHAR2(30)
GENDER CHAR(1)
On the EMP table, I have an index on the ID column and an index on the
GENDER column. Since the ID is unique, querying for a specific ID can
be most often facilitated by using the index. The Rule Based Optimizer
(RBO) would always use this index to perform the query. Now let's query
the same table for all females (GENDER='F'). The RBO, using the same
rule, would use the index on this column to execute the query. Assuming
equal distribution, the same number of men and women in the company, an
index would be less effecient than simply scanning the entire table.
Let's further assume that this table only has five rows of data that all
fit comfortably in one database block, which requires only one I/O
operation to read. Reading an index block, even for the ID column,
would be more inefficient than performing a full table scan. SQL
statement tuning when using RBO came down to the DBA obtaining an
understanding of the data distribution and determining the best access
path not found according to the RBO. The DBA would then rewrite the SQL
statement to use the rule the DBA determined was most efficient.
The RBO has its share of problems as the above examples show. The RBO
did not take into account data distribution or the size of the table.
The Cost Based Optimizer (CBO) uses this type of information to make
better informed decisions. The CBO needs statistics to be able to make
informed decisions on the quickest access paths. SQL statement tuning
under the CBO involves making sure that the correct statistics are in
place, and even locking down statistics so that changes to the data do
not affect the SQL statement processing. Additionally, the DBA can make
use of Stored Outlines so as to get the same consistent access path for
the queries. The DBA can still use SQL hints to influence the CBO as
well, similar to the RBO.
|
