Choice of Optimizer

If the reader will now use rule-based optimizer, by all means, it is the transition to cost-based optimizer. Reasons for this are many.


  • The time spent on coding is reduced.
  • The programmer does not need to know the rules.
  • Optimizer has more features, and many more tools for tuning.
  • Significantly increased the chance of finding packages from different suppliers, but which operate efficiently. Many of these packages developed for platforms running on DB2, Informix, SQL * Server, and Oracle. The code is written so as to meet the needs of the rule-based optimizer, but is a general.
  • End users can develop tuned code without having to learn a sizeable resource optimizer rules.
  • Optimizer cost is gradually improved in future versions of Oracle. The development of rule-based optimizer has been abandoned.
  • There is less risk associated with adding new indexes.
  • There are many features that are available only in the case of cost-based optimizer. These include the identification of prospects materialized (materialized views), the transformation of a star (star transformation), the use of functional indexes, etc. The number of similar mechanisms is huge and over time the differences between optymalizatorami will grow.
  • Oracle has made a number of mechanisms (such as DBMS_STATS package and scenarios) to resolve known problems with the inconsistencies of cost-based optimizer in different environments.

No comments: