Skip to main content

Overview

An ontology defines dataset rows as entities such as customers, products, and orders and connects relationships between them. Use an ontology when you need to query multiple datasets with the concepts and names used by the business.

Modeling list with entity and relationship counts by collection

Core concepts​

ConceptDescriptionExample
EntityA business object managed as a distinct type.Customer, product, order
RelationshipDefines how two entities are connected.Customer places order, order contains product
PropertyA data field on an entity or relationship.Customer ID, order date, quantity
Identifier keyA column that distinguishes each data row.customer_id, order_id
Display columnA column shown as the representative name in lists and graphs.Customer name, product name
Source datasetA dataset that supplies properties and actual values for an entity or relationship.Customer table, order table

Each collection is associated with one ontology scope. Manage entities and relationships under Modeling and Graph Explorer in the sidebar, rather than as leaf items in the collection tree.

Create an ontology​

  1. Select the collection to work in from Ontology Builder.
  2. Create an entity and set its name, properties, identifier key, and display column.
  3. Create relationships between entities and set their direction and properties.
  4. Map entity and relationship properties to dataset columns in the same collection.
  5. Load data with a pipeline, then review rows on the Data tab.
  6. In Graph Explorer, select an entity or relationship to query actual connections.
Review the schema before creation

After creating an entity or relationship, you cannot directly change an existing property's name, type, or nullability, or change the identifier key. Compare the source dataset columns with the identifier key before creation.

Review data​

Select Data in an entity or relationship row in the Modeling list to review rows from its mapped dataset. The display column appears as the representative name and the identifier key distinguishes each row.

In Graph Explorer, select an entity or relationship type to query connected data. Select a result node or edge to review its properties.

If source changes do not appear on the ontology screen, ask an administrator to review synchronization status. An administrator can run synchronization under Ontology mirror maintenance in Settings.

Continue​