Gold entities in Medallion architecture

Gold entities are a key component of the Medallion architecture, designed to deliver high-quality, business-ready data for analytics and reporting. After data is ingested and refined through the bronze and silver layers, the gold layer applies business logic using SQL views. These views define a centralized data model that standardizes reporting and analytics across your organization.

Persisting gold entities as tables improves performance and enables seamless integration with Power BI datasets, ensuring data is accessible for business intelligence and advanced analytics.

NOTE
Gold entities provide accurate, trusted data that is ready for end-user consumption and strategic applications. By applying business logic through customizable queries, the gold layer prepares data for analytics and reporting.

Centralized data model in the gold layer

The gold layer contains a centralized data model composed of dimension and fact views:

• Dimensions: Represent business attributes, such as customer details or product information.
• Facts: Capture measurable events, including sales transactions or quantities.

Create a gold entity

Use the gold entities panel in your workspace to create a new gold entity. This process enables you to define business logic and prepare data for downstream analytics and reporting.

IMPORTANT
When you add a gold entity in the NCC portal, the referenced view must exist. Otherwise, errors will occur in the pipeline.

Step 1: Access the Gold Layer in the NCC Portal

1. Navigate to Gold Layer.
2. Select + New Entity.

Step 2: Configure the gold entity

After selecting + New Entity, the Add Gold Entity window opens. Configure the following options:

Source

• Pipeline
  Choose one of the following options:

  • PL_GLD_COPY_FROM_SILVER
    Copies the view to the Warehouse for persistence.
  • PL_GLD_NOTEBOOK_FROM_SILVER
    This option is currently in development.

  • PL_GLD_PERSIST_FROM_SILVER
    Persists the view using a stored procedure in the Warehouse.

    IMPORTANT
    While the pipeline names include _FROM_SILVER, you must specify the actual source in the next configuration step.
    Selecting your Gold Warehouse sets it as the source, not Silver.

• Lake house
  Select either your Silver Layer Lakehouse or Gold Layer Warehouse as the source where the view is stored.

• Source schema
  Enter the schema containing your view.

• Source name
  Enter the name of your view.

Target

• Lake house
  Select your Gold Layer Warehouse.

• Schema
  Enter the schema for the persisted table.

• Source name
  Enter the desired name for the persisted table.

• Primary keys
  Configure primary keys if your view includes them. (Optional)

• Incremental
  This option is currently in development.

• Load order
  Specify the load order for persisting tables, especially when dependencies exist between layers. Assign a numeric load order to each table to ensure dependencies are respected.

Example Load order

For example, if a fact table depends on dimension tables (such as Customer or Product), ensure those dimension tables are loaded first.
This guarantees that all required reference data is available before loading the dependent fact table (e.g., SalesTransactions).

graph TD
    Customer["Dimension: Customer <br> Load order: 1"]
    Product["Dimension: Product <br> Load order: 1"]
    SalesTransactions["Fact: SalesTransactions <br> Load order: 2"]

    Customer --> SalesTransactions
    Product --> SalesTransactions

In this diagram, the SalesTransactions fact table (load order 2) depends on the Customer (load order 1) and Product (load order 1) dimension tables. Load the dimension tables first to satisfy dependencies.