Model Relational Data as a Cube

With Essbase on Autonomous AI Lakehouse, you can organize relational data as a multidimensional cube that reflects how your business analyzes information.

At this modeling stage, you can think of the process as cubing your relational data: in other words, organizing relational business data into dimensions, hierarchies, measures, and calculation relationships that reflect how your users need to analyze the business.

Suppose your relational data contains sales records with columns for product, market, date, revenue, cost, and units. In SQL, you can query and aggregate these values directly. In Essbase, you model the same business domain using dimensions, members, hierarchies, measures, and calculation relationships.

For example:

Relational data	Essbase model
Product identifiers and classifications	Product dimension
Country, region, and market values	Market dimension
Dates and reporting periods	Time dimension
Revenue, cost, and units	Measures
Business rollups	Member hierarchies
Derived business values	Calculation logic

The resulting cube gives analysts a consistent, business-oriented structure for analyzing the data.

A dimension represents a major perspective on the business, such as Product, Market, or Time. Members represent the individual values within each dimension. Hierarchies organize those members into parent-child relationships, such as:

Total Market → Americas → United States → California

Measures represent the numeric values to analyze, such as Revenue, Cost, or Units.

Essbase can also add calculation logic that gives those values additional business meaning. For example, a model can define Gross Margin from revenue and cost, or define how lower-level members contribute to higher-level totals.

Together, these elements form the Essbase outline: the multidimensional structure through which users of your cube analyze the relational business data.

When designing the cube, focus on how end users need to view and analyze the business data, rather than attempting to reproduce the physical relational schema.

After you establish the outline model, you can further shape its dimensions and hierarchies, add business logic with calculations, build the cube using an application workbook, and analyze it using Essbase query and analysis tools.