Designing a Database › Technical Design Defaults › Transformation

Transformation

Transformation is the process of creating a data structure diagram (DSD) based on the ERD. You access the DSD through the combination of the Data Structure List and Data Store List Tools.

The DSD consists of data structure objects (tables, indexes) and data store objects (databases, tablespaces, indexspaces). Based on the data objects of the DSD, the DDL is created during Generation.

Before performing transformation, CA Gen runs a consistency check against the data model. If any portion of the data model is incomplete or inconsistent, the transformation fails. The design team can either correct the errors in the data model and then transform again, or manually transform parts of the data model that were consistent. Manual transformation involves using the Data Structure List tool and requires more work from you than when CA Gen performs transformation.

CA Gen transforms a data model in the following way:

• Each entity type in the data model is implemented as a table.
• Each attribute in the data model is implemented as a column.
• Each relationship with a cardinality of 1:1 or 1:M is implemented as a foreign key. A foreign key is a collection of one or more columns in one table that is used to identify another table.
• Identifiers are implemented as indexes.
• Foreign keys (implementing relationships) are placed in indexes.
• Data set sizes are calculated from entity type and relationship properties (maximum number of occurrences, and so forth).
• Names of tables and columns are determined in this way:
  • If the entity type (being implemented as a table) or attribute (being implemented as a column) has a Data Structure List name, it is used.
  • If there is no Data Structure List name, the entity type or attribute name is used.
  • If the resulting name is a reserved word in the selected database management system, the name is modified slightly by CA Gen.
  • All tables are placed in a single database.

  Note: The characters in the table name are not verified for validity on the target DBMS OS system. The designer must ensure that any invalid characters generated in table or row names are replaced with valid character sequences. This can be done by defining a valid sequence of characters for the TD name of the entities and attributes or by replacing the generated TD names in the Data Structure List with valid characters sequences.