Introspection is the process of collecting native data source metadata so that all selected resources—like catalogs, schemas, tables, and procedures—are represented in a standardized way in the TDV virtual data modeling layer. The representations include capabilities and connection profiles. Composite views and procedures can then be built on the data modeling layer.

Introspection can be performed any time after data source configuration. When setting up introspection, you can also set filter criteria so that any new resources of interest are automatically introspected.

Introspected resources are added to the Studio resource tree for use in the Studio modeling layer, where you can then create views, procedures, and other resources based on those data resources. The metadata layer can be further enhanced by definition of indexes and primary keys, by discovery of data relationships, and by gathering of statistics about the data. The TDV query engine uses introspection data to generate efficient queries with source-aware algorithms.

See Retrieving Data Source Metadata for more information.

After each data source instance is configured and introspected in Studio, you can use that instance with any other data source instance to create views and procedures that intermix them all.

Note: Privileges on objects in the Data Source are evaluated only at the time of introspection or later when a FETCH is called during statement execution. So if privileges on an object in the data source are lowered after introspection, the error will not be caught until actual statement execution.

Reintrospection is the process of updating the data modeling layer after a data source has changed. During reintrospection, new data resources can also be automatically introspected for the first time.

Reintrospection can be initiated in any of these ways:

You can set filters that automatically reintrospect based on criteria applied to a catalog, schema, node, or resource.

Introspection and reintrospection can be performed in the background. Studio provides status reports about the resources selected. Multiple introspection and reintrospection tasks can be initiated on many resources. After resources are introspected, reintrospection is needed only if resource metadata has changed.

The introspection steps are similar for all data sources. They fall into three phases, each of which affects performance: