![]() For example, it is common to see Dim and Fact used as prefixes of tables in a relational star schema. You should remove any prefix and any suffix you might use in table names. The names of both views and columns exposed in the views should be user friendly and identical to names exposed to the users. ![]() However, adding the relationships manually to the data model is only a minimal cost Use meaningful names This is important when you want to include business logic that should be shared across different data models, so you do not have to duplicate the same transformation logic in several Power BI data models.īy importing views instead of tables, the data model might not recognize all the existing relationships between tables, because referential integrity constraints are applied to tables and not to views. The views can include transformation of data. The views created should include an explicit list of columns, and should not be a generic one such as: You just have to deploy a compatible version of the views that use the new structure, and notify to the BI analysts who owns the data model that they might use a new version of the data, coordinating with them how to provide the new structure (for example, by changing existing views or by providing different views). You will not delay the deployment in production of certain database changes, because you do not have to synchronize the deployment of a new version of all the existing reports. If you modified the database structure, probably you want to reflect this change to the reports, but with a different timing. Changing the view to keep compatibility with existing reports is usually a first temporary step. Keeping all the views for a data model in the same schema simplifies the tracking of the dependent reports. For example, in SQL Server you can use built-in features (such as View the Dependencies of a Table) or third party tools (such as SQL Dependency Tracker from Red Gate). Moreover, it is much easier to track dependencies between views and tables in a single relational database. Before publishing in production a change in the database structure, it is possible to adapt these views so that they will continue to work returning the same content, without breaking the refresh of existing reports.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |