The Semantic Modeler in Oracle Analytics Cloud

October 11, 2022 | 5 minute read
Pravin Janardanam
Director, Product Management, Oracle Analytics
Text Size 100%:

Oracle Analytics has a seasoned, rich Semantic Model, which has been used by thousands of analytics customers over the past two decades. Today, I'm excited to share with you the next generation modeling tool and a modeling language to create those semantic models.

Semantic Modeler

Data Model Tools Today

Let's take a quick look at the modeling tools available in Oracle Analytics today:

Semantic Modeler (New): A web-based tool for creating semantic models and publishing the semantic model as an RPD file for deployment. We'll cover this in detail in this post.

Model Administration Tool: A developer-focused modeling tool that provides complete governed data modeling capabilities by enabling a developer to define rich business semantics, data governance, and data interaction rules to fetch, process, and present data at different granularity from disparate data systems. The semantic model created by a developer using this tool is the brains behind the intelligent analytics query engine that generates optimized queries across various data sources.

Data Modeler: A web-based, easy-to-use data modeling tool to create simple semantic models. It doesn't cover the complete spectrum of semantic model capabilities. Data Modeler doesn't support creating semantic models that involve data at different grains and federated and fragmented data sources. Data Modeler will be replaced by Semantic Modeler.

Dataset Editor: A simple, easy-to-use, data modeling and data preparation tool that empowers data analysts and business analysts to create datasets based on data from local and remote files, including more than 50 connections and Subject Areas. Datasets enable business users to create self-service data models on top of existing governed semantic models.

Web-based Semantic Modeling Tool

The new Semantic Modeler is web-based, with a modern interface for data modeling, and it provides a streamlined user experience to create governed semantic models. It has a tight integration with Git to provide a seamless multi-user development experience. Developers create the models using the Semantic Modeler UI, or they can create the models using the Semantic Model Markup Language (SMML). 

Semantic Modeler is an alternative to the Model Administration Tool with complete semantic modeling capabilities currently available in Oracle Analytics Cloud for relational sources only. Semantic Modeler generates Semantic Model Markup Language (SMML) to define semantic models. Developers fond of creating semantic models with code can use SMML directly. Those who prefer to use a modeling tool with diagramming capabilities to build a model can use Semantic Modeler.

Salient features of Semantic Modeler include:

  • Modern and browser-based modeling tool that's an integrated component of Oracle Analytics Cloud.
  • Complete semantic modeling capabilities including physical diagrams, logical diagrams, and lineage diagrams.
  • Tight integration with any Git-based platform. You can perform most common Git operations from within Semantic Modeler.
  • Transparent SMML generation to define semantic models.
  • An SMML editor that includes smart integration with an expression editor to validate calculations and advanced expressions.
  • A lineage viewer that shows the mapping of physical, logical, and presentation layers.
  • Streamlined search integration that seamlessly shows relationships among objects.


Here is a sneak peek of Semantic Modeler features:

Physical Diagram


Logical Diagram


Git Integration - View Diff


Git Integration - Merge Conflict

Language to Define Semantic Models

Introducing Semantic Model Markup Language (SMML), a modeling language based on the popular JSON format familiar among developers to add business semantics to data. SMML provides a grammar, syntax, and structure for defining semantic models. A semantic model is a collection of SMML files that can be version controlled with the Git integration.


SMML files organized by Semantic Layers

  • Design-time semantic model definitions in human-readable format with SMML can be exported to a run-time RPD file for deployment.
  • Each semantic model object is a file containing SMML definition.
  • Files names map to what you see in the UI for object names.
  • File granularity is at the table level, thereby reducing the number of files to manage.
  • Object references are easy to define with fully qualified names of the objects.
  • SMML (along with Git integration) enables collaborative multi-user development and version control of semantic models.


SMML Editor in Semantic Modeler

Preview and share your feedback

The current semantic model supports only relational sources. If you have a semantic model with Oracle Essbase, Oracle OLAP, or Analytics Views, continue to use the Model Administration Tool rather than Semantic Modeler. 

To preview Semantic Modeler in your environment after the May update, display the Console and select System Settings and Preview to enable the Semantic Modeler. You must have Service Administrator privileges to perform this action. 

Animated GIF

Enable Semantic Modeler Preview

Try importing your existing semantic models (RPD files) or create a semantic model from scratch. Share your experience with us.


For additional information, se the  Semantic Modeler Guide and Semantic Model Markup Language reference.

Learn more about Oracle Analytics. Follow us on Twitter@OracleAnalytics, and connect with us on LinkedIn.

Pravin Janardanam

Director, Product Management, Oracle Analytics

Previous Post

OAC maps in Classic (Answers) can now source their metadata from ADW

Next Post

Begin every data journey with Auto Insights

Al Walker | 3 min read