Transforms › Refine a Model Using Transforms

Refine a Model Using Transforms

A logical model and physical model represent design layers in the application development process. In each design layer, you can perform design activities that correspond to steps and decisions in the development process.

After you create a logical or physical model, you can automatically change model objects using a transform. A transform is a method by which you can apply a design layer decision and maintain a record of the design decision. You can apply transforms to model objects using wizards to refine and optimize a design layer. For example, you can refine the physical model to improve query performance by applying a transform to two tables and creating a single denormalized table.

You can also apply transforms consecutively, to create a nested transform. Transforms can only be nested in a sequential order. If you want to add a new transform that would be based on some objects already involved in another transform, then the first transform has to be in target mode. The target objects of the first transform become the source objects for the next transform.

Important: When you attempt to apply a new transform on some source objects of an existing transform then the corresponding toolbar buttons are disabled. However, if you switch the first transform to be showing targets then the toolbar buttons become enabled as the new transform is now allowed.

The main advantages of using transforms include:

• Automation

  Simplify the refinement of the logical and physical model. You can use wizards to automatically apply design layer changes, instead of manually applying changes.

• Traceability

  Maintain historical information for each model object that the transform creates. You can trace the history of transformed objects.

• Preservation of object properties

  Preserve the properties of the transformed objects. You do not have to manually reenter the information.

You can use transforms to refine a model by:

• Rolling down two tables or entities and create a single denormalized table or entity
• Rolling up two tables or entities and create a single denormalized table or entity
• Resolving many-to-many relationships
• Rolling up subtype relationships
• Rolling down subtype relationships
• Creating an identifying relationship between a supertype entity and its subtype entities.
• Horizontally partition an entity or table
• Vertically partitioning and entity or table
• Copying and linking columns (column denormalization)