Domain 5: Autonomous Database Tools (11%)

Domain 5 of the 1Z0-931-25 Oracle AI Autonomous Database 2025 Professional exam covers the integrated tools available within Autonomous Database for machine learning, application development, data integration, analytics, and database administration. This domain represents approximately 6 questions on the 50-question exam (90 minutes, 68% passing score).

The exam syllabus defines six objectives for this domain (Exam Syllabus):

1. Describe Autonomous Database Tools
2. Use Autonomous Database with Oracle Machine Learning, APEX and SQL Developer Web
3. Explain Data Transforms
4. Create Business Model
5. Apply Data Insights
6. Explain Data Analysis

1. Autonomous Database Tools Overview

Every Autonomous Database instance ships with a suite of integrated, browser-based tools accessible through Database Actions (formerly SQL Developer Web). No separate installation, licensing, or provisioning is required. These tools are grouped into functional categories on the Database Actions launchpad. (Oracle Database Actions)

Exam trap: Database Actions and APEX are accessed through separate URLs. Database Actions is launched from the ADB console "Database Actions" button. APEX is launched from the "Tools" tab on the ADB details page. They are distinct entry points despite both being web-based.

2. Oracle Machine Learning (OML)

Oracle Machine Learning provides in-database machine learning on Autonomous Database. Algorithms execute inside the database engine against the data where it resides, eliminating the need to extract data to external ML platforms. (OML Documentation)

OML Components on Autonomous Database

Source: OML Product Index

AutoML UI in Detail

AutoML UI automates the full ML pipeline through five stages (AutoML UI Documentation):

1. Algorithm Selection -- ranks algorithms by likelihood of producing the most accurate model for the given dataset
2. Adaptive Sampling -- finds an optimal data sample size to accelerate subsequent stages
3. Feature Selection -- identifies the most predictive subset of features
4. Model Tuning -- optimizes hyperparameters for shortlisted algorithms
5. Feature Prediction Impact -- computes the influence of each input column on predictions

Supported task types and algorithms:

Key AutoML settings:

• Maximum Top Models: 2-5 (default 5)
• Maximum Run Duration: Up to 8 hours (default)
• Database Service Level: Low (no parallelism, more concurrent jobs), Medium, or High (max parallelism, fewer concurrent jobs)
• Execution mode: "Faster Results" (smaller hyperparameter search) or "Better Accuracy" (broader search, longer runtime)

The Leader Board displays top-performing models ranked by the selected metric. From it you can deploy a model to OML Services, rename it, or generate an OML4Py notebook that recreates the model. (AutoML UI Documentation)

Exam trap: AutoML UI supports only classification and regression. It does not support clustering, anomaly detection, or association rules. If a question asks about unsupervised learning in AutoML UI, the answer is that it is not supported there -- use OML4SQL or OML4Py instead.

OML Services REST API

OML Services provides REST endpoints for model deployment and scoring without provisioning separate infrastructure. Key capabilities (OML Services Documentation):

• Deploy both in-database OML models and third-party ONNX format models
• Real-time scoring via REST endpoints
• Model lifecycle management (versioning, activation, deactivation)
• Cognitive text analytics: topic discovery, keyword extraction, text summarization, sentiment analysis (English only), and feature extraction. Supported languages: English, Spanish, French, Italian.
• Cost advantage: charges only for actual scoring, no always-on VM required

Exam trap: OML Services cognitive text capabilities include sentiment analysis only in English. Topic discovery, keywords, summary, and feature extraction support English, Spanish, French, and Italian.

3. Oracle APEX on Autonomous Database

Oracle APEX is a low-code development platform included with every Autonomous Database instance. It enables rapid application development through a declarative, browser-based environment. (APEX on ADB)

Key Facts

• Each ADB instance includes a dedicated APEX instance (not shared across databases)
• No precreated workspaces exist at provisioning time; the ADMIN user must create workspaces via APEX Administration Services
• Only the ADMIN database user has initial APEX access; other users need workspace provisioning
• APEX versions are automatically maintained by Oracle with patches and upgrades
• Upgrade window: 45 days by default (extendable to 90 days) to apply new APEX releases
• Accessed via the Tools tab on the ADB details page, not through Database Actions

APEX Capabilities on ADB

APEX Service Workload Type

The exam distinguishes between APEX running on any ADB workload type versus the dedicated APEX Service workload type. APEX Service is a specialized, lower-cost ADB workload optimized specifically for APEX application development. It can be promoted to ATP if requirements grow beyond APEX-only usage.

Exam trap: APEX is available on all ADB workload types (Lakehouse, ATP, JSON Database, and APEX Service). The APEX Service workload type is simply the most cost-effective option when APEX is the sole use case.

4. Database Actions (SQL Developer Web)

Database Actions is the web-based administration and development interface for Autonomous Database. It is powered by Oracle REST Data Services (ORDS) and requires only a browser. (Oracle Database Actions)

Core Development Tools

Administration Tools

Exam trap: Database Actions requires database user authentication, not OCI IAM authentication. Users log in with their database credentials. Access to specific tools is determined by database roles and privileges granted to that user.

5. Data Studio Tools

Data Studio is the integrated data management suite within Database Actions. It unifies data loading, transformation, modeling, analysis, and sharing in a single no-code environment. (Data Studio Features)

Data Load

Ingests data from multiple sources into ADB:

• Local files, spreadsheets, and object stores (OCI, AWS, Azure, Google Cloud)
• Automatic schema and column type detection
• AI-enhanced natural language commands for data engineering tasks (cleaning, transforming)

Data Transforms

A full-featured, built-in data integration tool for data engineers (Data Studio Features):

• 100+ prebuilt connectors via drag-and-drop UI
• Full and incremental data extraction with scheduled job execution
• Data restructuring, preparation, cleaning, and enrichment
• Generative AI pipeline automation for generating and storing vector embeddings
• Data Transforms can invoke Data Studio Data Load jobs within a Data Flow workflow

Exam trap: Data Transforms is an in-database integration tool, not a separate OCI service. It runs inside the Autonomous Database environment. Do not confuse it with OCI Data Integration (a separate OCI service for ETL across cloud resources).

Business Models

Business Models establish a semantic layer over your database tables (Data Studio Features):

• Define business entities with hierarchies, dimensions, measures, and calculations
• Create a shared model that multiple analysis applications consume
• Query performance automatically optimized by ADB
• Models are reusable across Data Analysis, Data Insights, and external tools

Data Analysis

Data Analysis provides analytical and business intelligence capabilities built on Analytic Views -- reusable SQL objects that define hierarchies, measures, and calculations over base tables (Data Studio Features):

Analytic Views are the core technical mechanism:

• An Analytic View is a database object (created with CREATE ANALYTIC VIEW) that declares hierarchies (e.g., Time: Year > Quarter > Month > Day), measures (e.g., Revenue, Units Sold), and calculated measures (e.g., Year-over-Year Growth)
• ADB pre-computes aggregations from Analytic Views, optimizing query performance automatically
• Analytic Views can be queried directly via standard SQL or consumed by external BI tools
• They serve as the data layer for both Data Analysis and Data Insights

Data Analysis capabilities:

• Build Analytic Views from Business Models or directly from base tables via the no-code UI
• Microsoft Excel and Google Sheets integration via add-in -- users query live ADB data natively in their spreadsheet without exporting CSVs
• Ad-hoc reporting and visualization on enterprise data through the browser interface
• Prepare Analytic Views for consumption by Data Insights, Oracle Analytics Cloud, and third-party BI tools

Exam trap: Data Analysis produces Analytic Views, not raw query results. If an exam question asks what Data Analysis creates or what differentiates it from running SQL directly, the answer is Analytic Views -- reusable, pre-aggregated, hierarchical data objects that optimize repeated analytical queries.

Data Insights

Data Insights provides automated anomaly detection and pattern discovery using built-in statistical and machine learning algorithms (Data Studio Features):

How Data Insights works:

1. Select a data source: a Business Model, an Analytic View, or a base table
2. Choose metric columns to analyze (revenue, counts, durations, etc.)
3. Run the analysis -- Data Insights applies multiple algorithm categories automatically:
   • Statistical outlier detection: Identifies values that deviate significantly from expected distributions
   • Trend analysis: Detects upward/downward trends, seasonality, and inflection points over time hierarchies
   • Correlation discovery: Finds unexpected relationships between dimensions and measures
4. Review ranked findings: results are presented as a prioritized list of insights, each with a natural language description, supporting visualization, and confidence score

Key characteristics:

• Fully no-code: no SQL, no algorithm selection, no parameter tuning required
• Operates on the semantic layer (Business Models / Analytic Views), not raw tables, so insights respect business definitions
• Findings can be exported, shared, or used to trigger further investigation in Data Analysis
• Runs inside the database -- no data movement to external analytics services

Exam trap: Data Insights is an automated discovery tool that surfaces patterns you did not ask for. It is not a query tool (that is Data Analysis) and not a machine learning model builder (that is OML). If a question asks which tool discovers anomalies without user-specified queries, the answer is Data Insights.

Catalog

Enterprise data discovery and governance:

• Unified view of data across clouds and on-premises, including Apache Iceberg tables
• Connects to external catalogs: Databricks Unity, Snowflake Polaris, AWS Glue
• Data lineage tracking showing dependencies and impact of changes

Data Sharing

Secure data distribution across organizations:

6. Tool Integration Map

Understanding how the tools connect is critical for exam questions about workflows:

Data Load --> Data Transforms --> Database Tables
                                       |
                                 Business Models
                                    /       \
                           Data Analysis   Data Insights
                                |
                          Analytic Views
                                |
                    Excel / Sheets / BI Tools

• Data Load brings data in from external sources
• Data Transforms cleans, enriches, and pipelines data
• Business Models define the semantic layer over tables
• Data Analysis creates analytic views and enables ad-hoc querying
• Data Insights runs automated discovery algorithms against models or tables
• Catalog provides governance and lineage across all stages

OML vs. Select AI: When to Use Which

This comparison sits at the Domain 5/Domain 6 boundary and is a likely exam scenario question. Select AI is covered in detail in the Domain 6 Study Guide.

Exam trap: OML and Select AI are complementary, not competing. OML builds predictive models; Select AI translates questions into SQL. A question asking "which tool enables business users to query data in natural language" expects Select AI, not OML.

Key Exam Traps Summary

1. Database Actions and APEX are separate entry points. Database Actions is launched from the ADB console. APEX is launched from the Tools tab.
2. AutoML UI supports classification and regression only. No clustering, anomaly detection, or association rules.
3. OML sentiment analysis is English-only. Other cognitive text features (topics, keywords, summary) support English, Spanish, French, Italian.
4. APEX is available on all ADB workload types, not just the APEX Service workload.
5. Data Transforms is in-database, not a separate OCI service. Do not confuse with OCI Data Integration.
6. Business Models create a semantic layer, not a machine learning model. The term "model" here means a logical data model with dimensions and measures.
7. Database Actions uses database authentication, not OCI IAM credentials.
8. No precreated APEX workspaces exist at provisioning. The ADMIN user must create them.
9. OML Services charges only for scoring, no always-on VM cost. This is a cost advantage over standalone ML deployment.
10. Graph Studio supports both property graphs and RDF graphs, with 80+ built-in algorithms.

References

• Oracle Machine Learning Documentation
• AutoML UI Documentation
• OML Services REST API
• APEX on Autonomous Database
• Oracle Database Actions
• Data Studio Features
• Graph Studio on Autonomous Database
• 1Z0-931-25 Exam Syllabus
• Oracle Education: 1Z0-931-25