Creating and accessing physical data models

Before you can design data flows and mining flows, you need to create physical data models for your source and target databases. These models contain the metadata for the tables and other database objects that you include in your flows.

In many cases the data warehousing architect and the data model designer are different people. Before you proceed, you need to know if any physical data models already exist in your enterprise, and if so, how they were created. Taking the time to understand the location and format of existing data models can save you work.

Based on this information, you can set up your environment in different ways:
  • If someone in your team already used the Design Studio to create physical data models from scratch (native "SQL models"), you can reference these models in your data warehousing projects. See Creating a data warehousing project.
  • If you want a physical data model to be created for you automatically based on a live database connection, select "online mode" when you create your data flows and mining flows. See Creating a data flow.
  • If physical data models already exist but they are in a proprietary format, such as ERwin, you can use the Design Studio to import these models and transform them to the SQL model format.
  • If no physical data models exist but you can connect to the live data sources via the Data Source Explorer or have access to the DDL scripts that were used to create the data sources, you can reverse-engineer the models that you need.


Feedback | Information roadmap