Import processes

Integration Composer has a Java™ based user interface application and an engine. The user interface is used to create and manage object models and rules for data transformation. The engine transforms source database instances and transfers them to the target according to a set of defined rules.

Classes from the source database are mapped to target class properties using the mapping rules. Properties from more than one source class can be mapped to a single target class or single target property. The mapping rules can be as simple as a one-to-one mapping or can involve a Java expression.

When Integration Composer runs a mapping, it reads source database instances, applies the rule set, and compares the transformed data against the associated target instances, if any exist. It then updates the target database, as follows:

• If the engine finds a new instance during the data transfer process, the engine inserts that instance into the target.
• If the engine encounters an existing target instance, it evaluates the instance to determine whether newer data exists. If new data exists, the engine updates the target instance. Otherwise, the instance is skipped.
• If data exists in the target but does not exist in the source, the instance might be deleted from the target. Deletion depends on qualifiers. For example, in the Deployed Assets 7.5 data schema, instances of deployed asset classes are not deleted, but instances of software classes can be deleted.

Most discovery agents store a data collection time stamp. For increased performance during mapping execution, Integration Composer compares the source time stamp for a top-level instance against an internally stored time stamp, a last scan time stamp. If time stamps differ, the top-level instance along with its child and reference class instances are processed, and the internally stored time stamp is updated. Otherwise these instances are skipped.

The diagram shown illustrates how the data is processed.