The only entities left to ask this question of are standalone entities and child-only entities having no optimal clustering.
You should choose CALC location mode if application programs always retrieve this entity using its full key or if it is relatively dynamic (that is, many additions, deletions, or key changes).
If an entity is relatively static and multiple occurrences are often retrieved together, it is most effective to cluster the entity through an index defined on the most-commonly used access key.
If the entity is not static, but often participates in multi-occurrence retrievals, cluster the entity on an index defined on its db-key. For more information on indexes, Chapter 12, "Refining the Database Design"
Copyright © 2014 CA.
All rights reserved.
|
|