The relationship discovery capability for catalog sources identifies pairs of similar columns and relationships between tables within a catalog source.
In order to discover relationships, the relationship discovery service uses predictions made by relationship inference models. The relationships that this service infers are referred to as Inferred Relationships. The relationship discovery capability allows you to infer the following relationships:
•Column Similarity: The identification of semantically similar pairs of columns within a catalog source
•Joinable Tables Relationships: The identification of tables that can be joined based on the column similarity predictions.
To enable the relationship discovery capability for catalog sources, you must specify appropriate values for related parameters while configuring a catalog source. For information about enabling this capability while creating a catalog source, see the Catalog Source Configuration help.
To discover similar columns and joinable table relationships, ensure that you configure the following options:
•On the Data Profiling and Quality tab, enable data profiling.
•In the Parameters, select Keep Signatures and Values as the run mode.
These configuration options enable you to retain values of the columns in the profiling results and discover relationships.
Using the relationship discovery capability
You can use the relationship discovery capability in several ways, including the following:
•Provide recommendations of tables that can be joined with other tables.
•Assist in migrating data to the cloud by analyzing related objects.