Creating custom intelligent attributes

Oracle Unity allows you to create custom intelligent attributes.

In this topic:

Introduction

Custom intelligent attributes enable you to enrich your data objects with new attributes to meet your business needs. Ensure to review the ready-to-use intelligent attributes before creating custom intelligent attributes.

When you are thinking and planning about the custom intelligent attributes to create, it's important to clearly understand the operation or calculation you are looking to design, and then identify the key attribute behind the operation or calculation.

Use cases

Custom intelligent attributes are created for business use cases the ready-to-use intelligent attributes do not already meet. These are some examples of custom intelligent attributes:

  • Most purchased brand

    Tracks the brand purchased most often.

  • Percent of money spent, returned

    The percent of money a customer spent on purchases, then returned.

  • Average purchased item value

    The average value of individual items purchased.

  • Percent of discounted items purchased

    The percentage of items purchased by a customer that are discounted.

Metadata requirements

Custom intelligent attributes rely on other metadata objects to define the operations and calculations for the custom intelligent attribute. Note these metadata requirements.

Data source views

The Data source view linked to the new intelligent attribute must meet the following requirements. If one of the below requirements are not met, the custom intelligent attribute will not be successfully created.

  • A unique dataSourceViewID.

  • A complete sourceTableQuery that surfaces the attributes needed by the Model object.

Models

The Model linked to the new intelligent attribute must meet the following requirements. If one of the below requirements are not met, the custom intelligent attribute will not be successfully created.

Property Required? Data type Description
modelId Yes String Unique identifier of the Model.
type Yes String The type of Model. Possible values include: Data Wrangling/ Customer Analytics / Sales / Product.
dataSourceViewId Yes String The Data Source View ID for the associated Data Source View.
cube Yes Object The associated cube for the model.
incrementalField Yes Object

Information about the attribute from a Data Source View to use to increment the data. The parameters atype, dsvId, and attributeName are required.

incrementalFieldFormat Yes Date/timestamp Date and time format to use for the incremental field. Must be "yyyyMMDD".

Explore your models to find if the key attribute is already expressed within a model. If it is, then review its dimensions and measures to see if the desired operation or calculation is already exposed. If the operation or calculation is already exposed, you can reuse the model. Otherwise if the operation or calculation is not exposed, you must create a cube and define new dimensions and measures.

Cubes

A cube is a component of a Model. The Cube linked to the new intelligent attribute must meet the following requirements. If one of the below requirements are not met, the custom intelligent attribute will not be successfully created.

Property Required? Data type Description Possible values
cubeId Yes String The Cube's unique identifier.  
modelId Yes String The ID of the Model linked to the Cube.  
status Yes String

The status of the Cube.

  • NEW: Cube has been published.

  • READY: Cube has been processed at least once, and contains queryable data.

state Yes String The state of the Cube.
  • ACTIVE: Cube is active and ready to publish.

  • PAUSED: Cube will be skipped for processing.

  • DELETED: Cube is set to be deleted.

type Yes String The type of Model.
  • Data Wrangling

  • Customer Analytics

  • Sales

  • Product

dataSourceViewId Yes String The Data Source View ID for the associated Data Source View.  
cube Yes Object The associated cube for the model.  
incrementalField Yes Object

Information about the attribute from a Data Source View to use to increment the data. The parameters atype, dsvId, and attributeName are required.

 
incrementalFieldFormat Yes Date/timestamp Date and time format to use for the incremental field. Must be "yyyyMMDD".  

Within Cubes, "measures" define additional information about the custom intelligent attribute.

You will need to use a Cube measure to create the intelligent attribute. You can use the Cube measure as is (.ReferenceAttribute) or use it with supported functions (.FunctionAttribute).

If the Cube measure you need doesn't currently exist for the Cube, then follow the steps below to adding a Cube measure. Otherwise skip this task.

If the measure you are adding can be added with data in data source view attribute, continue with adding the measure to the Cube. If the attribute needed is not available in the data source view, follow the steps below for updating the data source view.

Creating custom intelligent attributes

Here are the high-level steps to create a custom intelligent attribute:

Step 1: Verify the attributeId is unique

You will need to verify that no existing intelligent attribute uses the attributeId you are creating.

To verify the attributeId is unique:

  1. Create an API call to retrieve the attributeId.
  2. Review the response to verify the attributeId is unique.

Step 2: Create the intelligent attribute

  1. Send a request to create an intelligent attribute, for more information see the examples.

  2. In Oracle Unity, visit the Intelligent attributes page to confirm the new intelligent attribute is listed. It will have an Inactive status.

Step 3: Activate the custom intelligent attribute

After creating the custom intelligent attribute, you will have to make it active so that you can access it from the segmentation canvas.

To activate the intelligent attribute:

  1. Open the navigation menu and select Intelligent attributes.
  2. Click the Action menu for the intelligent attribute and select Make active.

You will now have to publish the changes to implement the update to the intelligent attribute.

Step 4: Publish the changes

When you publish the changes, Oracle Unity adds the intelligent attribute to the MasterCustomer data object.

To publish the changes:

  1. Open the navigation menu and select Data model.
  2. In the top-right corner, click Publish changes.
  3. Note: If a publish job is currently in progress, you will receive a notification that the publish job cannot run until the publish job already in progress completes.

  4. Confirm the publish job and click Publish.
  5. A notification will display to confirm the publish job successfully started.

You can monitor the status of the job from the Jobs dashboard.

Step 5: Run the Identity resolution pipeline job

The last task is to run the Identity resolution pipeline job so that Oracle Unity can calculate the values for the intelligent attribute.

To run the Identity resolution pipeline job:

  1. Open the navigation menu and select Data feeds. The jobs will be displayed at the top of the list of data feeds with the other system jobs.
  2. Click the Action menu for System - Identity resolution pipeline job and select Start.

The custom intelligent attribute will now have values and be available on the segmentation canvas.

Learn more

Intelligent attributes

intelligent attribute, intelligent attributes, how to create intelligent attributes