Merge Your External Applications

If you have an external semantic model from standalone Oracle Analytics Cloud, Oracle Analytics Server, or Oracle Business Intelligence Enterprise Edition, you can merge it with the semantic model of Oracle Fusion Data Intelligence only in your non-production instance.

You can also merge your existing catalogs in the external Oracle Analytics Cloud with those in the Oracle Analytics Cloud instance associated with your Oracle Fusion Data Intelligence instance. Both external semantic model and semantic model extensions can co-exist and you can include the external semantic model in your semantic model extensions. On the Publish Model page, select "Yes" if you want the external semantic model to be included.

If the external semantic model references application roles and groups, then you must create them manually. The merge capability imports only the mappings and doesn't import the roles and groups definitions. Create the applicable groups in your identity provider, see Create Groups or Create a Group as applicable. Create application roles using the Security page in Oracle Fusion Data Intelligence, see Create an Application Role.

Merging external semantic models may require you to scale up your Oracle Analytics Cloud compute resources to ensure better performance. In particular, if you're merging a semantic model that was in use on a live Oracle Analytics Cloud instance, then Oracle recommends scaling up Oracle Fusion Data Intelligence's Oracle Analytics Cloud compute resources by the same number of CPUs as the instance the semantic model is merged from.

To conform common dimensions between Oracle Fusion Data Intelligence and your other Oracle Analytics Cloud, ensure that the:
  • Data types of the conforming columns are matching, else you can't merge.
  • Data for the mapped columns match, else queries don't fetch correct results.
  • Level of details at which data is stored in the dimensions match, else queries return incorrect results.
  • External semantic model is up to 25 MB in size.
  • Catalog objects are in folder names that don't conflict with the prebuilt folders of Oracle Fusion Data Intelligence and in /Shared/Custom/ folder.
  • New logical table source that you add has a higher priority group than the prebuilt logical table source. For example, if the logical table source of the prebuilt object is 2, then you must specify a priority group of 3 or higher.
  • Additional hierarchy has its own total level.

The external semantic model connects only to the Oracle Autonomous Data Warehouse associated with Oracle Fusion Data Intelligence. After enabling the external application merge functionality, if you make semantic model changes with data augmentation, then it'll result in unexpected errors. If such changes exist, then you must update the data augmentations to remove the semantic model changes. Ensure that you make the semantic model changes in the external semantic model and that you don't use system variables in your extensions.You must also ensure that the functional areas enabled in the target environment and the development environment from where the template semantic model was extracted are the same to avoid failure of the external application merge.

To successfully merge an external semantic model, you must complete these tasks in the sequence they are listed:
  1. Create a custom database schema in the Oracle Autonomous Data Warehouse associated with your Oracle Fusion Data Intelligence instance.
  2. Migrate or create database objects such as tables, views, and synonyms in the new schema.
  3. Grant semantic model access to the database objects.
  4. Configure Oracle Analytics Client Tools (Oracle BI Administration Tool).
  5. Export the semantic model template.
  6. Bring in the custom database objects into the physical layer of the semantic model template by either importing, or merging, or copying and pasting.
  7. Model the custom semantic model objects.
  8. Merge the external semantic model with the semantic model of your Oracle Fusion Data Intelligence instance.
  9. Validate the results.
  10. Bundle the merged semantic model to your production instance.