A data validation view is a virtual view generated from one or more data sources that you want to test. Create a data validation view when you need to customize the set of data before you use it in a table object. After you create a data validation view, you can add it to a table object. You can also add some views to another view.
You can create the following types of data validation views:
A view based on a single data source or another data validation view. You can create an aggregate view when you need to validate aggregate data from a nonrelational source, such as a flat file. When you create the view, you choose the group by columns, and you configure expressions to create the aggregated data.
For example, your company recently acquired another company, and you need to verify the average pay rate. You received a flat file that contains employee information. You create an aggregate view to group the data by employee pay grade. Then you run a test to identify the average pay rate of employees in each grade.
A view based on two or more heterogeneous data sources joined by key columns. Create a join view to validate data across data sources and to check for possible discrepancies. You can use a join view in an aggregate view or in a table object.
For example, you want to verify that all customer records that you migrated from multiple sources were moved to the target. You create a join view of the sources. You add the view and the target table to a table pair object. Finally, you create a test to verify that the customers in all the sources also exist in the target.
A view based on an SQL statement. The view can include multiple data sources and multiple calculations to produce a single data source. You can create an SQL view based on one or more relational tables. If you create a view based on more than one table, all tables must be in the same database and use the same connection object. You can use an SQL view in an aggregate view, a join view, or a table object.
For example, you want a combined result set of two data sources. You create an SQL view that contains a UNION operator.
A view based on source and lookup data that you want to test for key values and orphan records. You can create a lookup view based on flat file and relational data sources. If the view contains two relational sources, the tables must be in the same database and use the same connection object. You can use a lookup view in a table object.
For example, you want to check the validity of contents in a target table against the contents in the source. You create a lookup view with the source table and the lookup table. You add the lookup view and the target table to a table pair object. Finally, you create a test to compare the IDs in the view and the target.