Customize

You can customize Fusion Data Intelligence to satisfy business requirements and increase reporting capabilities.

Oracle recommends following the phased implemention approach and perform the customizations after the two prebuilt content rollouts. Performing customizations before signing off on the first two prebuilt content phases adds unnecessary risks to an implementation project. You must perform customization activities in an environment that you consider the master development environment.

You can extend Fusion Data Intelligence to modify and create reporting content, data content, semantic model content, and security assignments. These types don't necessarily depend on each other, aren't necessarily sequential, and may not all apply to the business requirements.

Refer to Customize Oracle Fusion Analytics for detailed guidance.

Custom Reports

Fusion Data Intelligence includes prebuilt reports and subject areas that allow users to create reports. As users create custom reports using the prebuilt subject areas, it's important to organize them and secure them adequately to ensure that the correct business users with the correct permissions are allowed access to the custom reports and projects.

There are multiple types of custom reports. Plan to save each type in specific folders with adequate permission.

Ensure that:

Developers or key power users define reports to be shared with specific groups of consumers.
An application role is available to identify all the consumers that must have read access to these reports.
A role is available for users who must have write access.
Shared folders with appropriate read or write access to store these reports for each role are available.
Service administrators always have full control for the folders.
Ad hoc reports created by business users for themselves are saved in the user’s folder.
Ad hoc reports created by a business user to be shared are stored in specific shared folders with write access to everyone. If needed, the user who creates the report can configure permissions to give access only to the selected users.

See:

This task requires Fusion Data Intelligence service administrator, Fusion Data Intelligence security administrator, and Fusion Data Intelligence authors.

Data Extensions

As a service administrator and data engineer, use data extensions to add additional data for analytics.

The storage capacity in Oracle Autonomous Data Warehouse is planned to support data coming from your subscribed Oracle Fusion Cloud Applications sources. However, if you add data coming from another source, you need to ensure that you have enough storage available. Fusion Data Intelligence includes 50 GB of storage for external data. If you're planning to load more volume, you need to add additional capacity to the database (in storage and potentially OCPU). Adding capacity will consume Universal Credits. Plan to extend data using any or all of these options in the same environment for different use cases:

Custom Data Configurations
Data Augmentation
Self-service using Oracle Analytics Cloud Datasets and Dataflows
Third-party ETL tool such as Oracle Data Integrator

See:

Custom Data Configurations

Custom data configurations are ready-to-use templates that enable you to bring data into the autonomous data warehouse quickly.

Currently the only template available is for Descriptive Flex Fields (also referred to as DFF). You only need to select the list of DFF columns you want to import for each object. Data is loaded in the pipeline managed by Oracle and added to the semantic layer automatically.

See Extend Data with Custom Applications.

Data Augmentation

Use data augmentation to add additional data from a supported data source that doesn't require transformation. Oracle Fusion Cloud Applications data must be available in a view object to be loaded through data augmentation. Data augmentation can extend existing dimensions, create new dimensions, and create new facts.

There are many benefits to using data augmentation:

Oracle manages the execution.
It's scheduled with the prebuilt pipelines, which is essential for data consistency (extract dates are the same).
There is no need to buy, learn, or manage another extraction tool.

But data augmentation has these limitations currently:

No data transformation.
No joining with other tables and data sources.
A limited number of supported data sources, most available in preview only.

Leverage views in the OAX_USER schema built on the augmentation tables to implement transformations and joins. Depending on the performance requirements, these views can be materialized. However, this is useful only for simple use cases. Performance and maintenance quickly become issues if the transformations are too complex (multiple joins, aggregates). In that case, using a third-party extraction tool is the best option.

See About Augmenting Your Data.

Self-Service using Fusion Data Intelligence Datasets and Dataflows

Fusion Data Intelligence allows users to upload datasets directly and use them for reporting or joining them with existing subject areas.

These datasets can come from a file or be defined using a connection to an external application. Datasets are designed to provide self-service extension capabilities to business users to bring in external data that's not in the semantic model. Dataflows transform and load datasets data, for instance, into a custom schema in the Fusion Data Intelligence database. Dataflows can load only a small volume of data and are currently accessible only by their owner. But once loaded in the database, the corresponding table can be added in the semantic layer. Depending on the mode selected, a dataflow can drop and recreate the target table when executed. Since tables used in the semantic layer must have access granted to user OAX$OAC, granting access by default to the entire custom schema is safer. Otherwise, execute the grant SQL statement after each execution of the data flow.

See:

Third-Party ETL (Extract, Transform, and Load) Tool

Fusion Data Intelligence allows users to upload datasets directly and use them for reporting or joining them with existing subject areas.

Any third-party extraction tool that supports Oracle Autonomous Data Warehouse can load external data into the Fusion Data Intelligence database. Oracle Data Integrator is an example of an extraction tool.

Plan to use an extraction tool to load external data as a last resort when none of the other solutions are appropriate. You must review data augmentation and connectors first and if you can't use either of these capabilities, then plan to use Oracle Data Integrator for Marketplace and other ETL tools.

There are a few elements that are important when loading data into the Fusion Data Intelligence database with an extraction tool:

Ensure that custom data loads don't run simultaneously with pipelines to avoid contention on database resources. You can determine that an incremental pipeline is finished when a new record is inserted into table DW_WH_REFRESH_SUMMARY with the process name _REFRESH_SCHEDULED.
Use the Low service connection to the autonomous data warehouse. Using the High or Medium service consumes too many resources and can cause performance issues for other processes running on the database.
Depending on the design of your load process, data inconsistencies and performance issues can appear in reports if users access the custom tables while they are being loaded. Try to load the data outside of business hours.

See Configuring Custom ETL from Fusion Data Intelligence.

Semantic Model Extensions

Plan to extend the semantic model to meet your business requirements. This entails creating and modifying subject areas to leverage the data extensions.

Some examples are:

Add or extend subject areas
Add or extend dimensions
Add facts
Add or extend hierarchies
Add session variables
Add derived columns

Explore these options to extend the semantic model:

Use data augmentation to automatically add new data into the semantic layer in some situations. When you use data augmentation is to load additional information, you can use it sometimes to extend the semantic layer. There are three different types of data augmentation:
- Extend Entity: Use this to add additional attributes to existing dimensions. Attributes loaded with this type of data augmentation are automatically in subject areas in the selected folder. This extension type is possible only with a join on the entity's primary key.
- Custom Dimension: Use this to create an entirely new dimension. Attributes loaded with this type of data augmentation are added in subject areas only if they are joined with a custom fact created with data augmentation.
- Custom Fact: Use this to create a new fact table. Attributes and measures loaded with this type of data augmentation are automatically added to the subject areas selected. Note that specifying a subject area is not mandatory.
A common usage of data augmentation is to extend a dimension with additional attributes, assuming that the join key is included in the data source.
Use the Semantic Model Extension wizards that can use tables created by the data augmentation processes. This is the primary tool to use to modify the semantic layer. Below are some examples, but they are not exhaustive:
- Data is loaded through data augmentation, but data augmentation cannot make the appropriate modifications in the semantic model. For instance, the join with the existing dimension isn't based on the primary key, or some transformations must be applied after loading the data.
- Extensions need to be added in prebuilt subject areas.
- Calculations are based on prebuilt facts.
A common usage of this tool is to create derived metrics based on existing fact metrics or to extend a dimension with attributes that can't be joined through data augmentation.

You can use them together in the same Fusion Data Intelligence environment. Though each possibility is helpful for specific use cases, review these recommendations:

When creating a custom model, a custom fact in particular, ensure that you join it with prebuilt dimensions so that prebuilt columns and custom columns are used in the same report.
Minimize the number of semantic layer customization steps as much as possible by applying all the required modifications to the corresponding object in each step. When additional customizations on the same object are added later, modify the existing step instead of adding a new one.

This task requires:

Fusion Data Intelligence Service Administrator
Data Engineer

See:

Security Extensions

You can use new and existing application roles to configure data and object security for the data and semantic model extensions.

As a service administrator, security administrator, and semantic modeler administrator, you can:

Provide data security by restricting user access to data using application data roles. This filters the data presented to users based on their application data roles. Data security configuration is significantly different depending on the source Oracle Fusion Cloud Applications. Each Fusion Data Intelligence application has its way of handling data security. Even though Fusion Data Intelligence platform features are always the same, they are used differently depending on the application. Security assignment configuration is fully, partially, or not automated, depending on the Fusion Data Intelligence application.
Set up object permissions to restrict user access to catalog content (reports and visualizations) and subject area content (subject area, folders, and columns) using the prebuilt and custom duty roles. Duty roles set either "Default (inherits from the parent element)", "No access", or "Read-only" permission levels to the objects they protect.
Customize the permissions for the prebuilt duty roles and create permissions for custom roles.

You must:

Assess the prebuilt data security configuration and determine what business requirements it satisfies. Use data security extensions for any requirements that aren't satisfied with the prebuilt data security configuration.
Document all security extensions.
Maintain matrices of the extensions and the corresponding application roles that secure them.
Validate the security extensions by testing them before publishing them to production.

See: