What’s New for Oracle Autonomous Database Serverless

May 2025

Feature Description

Feature	Description
Select AI Enhancements	Select AI includes the following new features: Select AI automatically detects relevant tables and send metadata only for those specific tables, relevant to the query, in Oracle Database 23ai. You can achieve this by specifying `object_list_mode` as automated. See "Automatically Detect Relevant Table Metadata" in Example: Improve SQL Query Generation. Select AI enables you to restrict table access by considering only the tables listed in the AI profile attributes for SQL generation. You can achieve this by specifying `enforce_object_list` to true. See Example: Restrict Table Access in AI Profile. Select AI enables you to specify case sensitivity such that the LLM produces case-insensitive responses from the database and LLM. You can achieve this by specifying `case_sensitive_values` to false. See Example: Specify Case Sensitivity for Columns. Administrators can control data access such that actual schema tables are not sent to the LLM using Select AI. See ENABLE_DATA_ACCESS Procedure, DISABLE_DATA_ACCESS Procedure, and Example: Enable or Disable Data Access.
Use In-database Transformer Model with Select AI RAG	You can use in-database transformers (also known as embedding models) that you imported into your Oracle Database 23ai instance with Select AI RAG to generate vector embeddings for document chunks and user prompts. See Use In-database Transformer Models and Example: Select AI with In-database Transformer Models to explore this feature.

Select AI Enhancements

Select AI includes the following new features:

Select AI automatically detects relevant tables and send metadata only for those specific tables, relevant to the query, in Oracle Database 23ai. You can achieve this by specifying object_list_mode as automated. See "Automatically Detect Relevant Table Metadata" in Example: Improve SQL Query Generation.
Select AI enables you to restrict table access by considering only the tables listed in the AI profile attributes for SQL generation. You can achieve this by specifying enforce_object_list to true. See Example: Restrict Table Access in AI Profile.
Select AI enables you to specify case sensitivity such that the LLM produces case-insensitive responses from the database and LLM. You can achieve this by specifying case_sensitive_values to false. See Example: Specify Case Sensitivity for Columns.
Administrators can control data access such that actual schema tables are not sent to the LLM using Select AI. See ENABLE_DATA_ACCESS Procedure, DISABLE_DATA_ACCESS Procedure, and Example: Enable or Disable Data Access.

Use In-database Transformer Model with Select AI RAG

You can use in-database transformers (also known as embedding models) that you imported into your Oracle Database 23ai instance with Select AI RAG to generate vector embeddings for document chunks and user prompts. See Use In-database Transformer Models and Example: Select AI with In-database Transformer Models to explore this feature.

Parent topic: What’s New for Oracle Autonomous Database Serverless

April 2025

Feature	Description
Data Studio Enhancements	The Data Studio is updated with new features and changes: The Autonomous Database Federated Database Catalog Enhancements: You can connect multiple catalogs to make it easier to find and work with your data, including popular Iceberg Data Catalogs like Unity and Polaris. You can also link to remote Autonomous Database instances using Database Links, allowing you to query tables from those databases directly within your catalog — helping you get more from your data catalog without extra steps. One of the major improvements in the Federated Catalog is the ability to easily load a Delta Share directly from the Catalog tool itself — no need to switch to a separate data loading tool. See Inline Data Load of Delta Share from the Catalog Tool for more details. The Oracle Autonomous Database for Excel Add-in can now be installed from the Microsoft Store. See Install the add-in from Microsoft AppSource for more details. UX Improvements for AI Table Assistant - The Table AI Assist tool includes the ability to edit recipes, has auto-save of recipes, UX improvements such as showing new column header in italics. See The Table AI Assist Tool for more details. Live Feed Reporting enhancements: Using the existing Simple Mail Transfer Protocol (SMTP) email registration setup in the Live Feed, you receive email when: New rows are added: You receive email notifications when Live Feed runs and loads your table with new data. Data Validation Issues Occur: You receive email notifications when inconsistencies arise, such as when the tool attempts to load non-numeric values into a numeric field. Live Feed throws an Error: You receive email notifications if errors occur such as when the bucket containing your files is deleted or your credential for that bucket was dropped. See Send Email using Live Feed Tool for more details.
Long Pending Transaction Event	The `LongPendingTransaction` event is generated when a transaction is open for longer than 24 hours. See Information Events on Autonomous Database for more information.
Recovery Time Objective (RTO) for a Cross-region Autonomous Data Guard Standby	The Recovery Time Objective (RTO) for cross-region Autonomous Data Guard Standby databases is lowered to less than ten (10) minutes. See Autonomous Data Guard Recovery Time Objective (RTO) and Recovery Point Objective (RPO) for more information.
Documentation Addition: Automatic Regression Detection	Automatic Regression Detection in Autonomous Database provides proactive handling of regressions and enables automated issue detection, diagnosis, and mitigation.  See About Scheduled Maintenance and Patching for more information.
Average Active Sessions and Session Utilization Metrics	The `AverageActiveSessions` metric shows the average number of sessions that are actively executing or waiting for resources. The `SessionUtilization` metric shows the maximum session utilization. See Available Metrics: oci_autonomous_database for more information.
Credential Comments in `DBMS_CLOUD.CREATE_CREDENTIAL`	The `DBMS_CLOUD.CREATE_CREDENTIAL` procedure supports comments when you use the `comments` parameter. See Create Credentials to Access Cloud Services and CREATE_CREDENTIAL Procedure for more information.
Migration from PostgreSQL, MySQL, or SQL Server to Oracle SQL	You can migrate SQL statements from PostgreSQL, MySQL, or SQL Server, to Oracle SQL and run the statements on Autonomous Database. See Migrate Applications from Non-Oracle SQL to Oracle SQL for more information.
Documentation Change: Pre-Authenticated Request (PAR) URLs Renamed to Table Hyperlinks	You can generate and manage Table Hyperlinks to access data on Autonomous Database. See Use Table Hyperlinks for Read Only Data Access on Autonomous Database for more information.
Additional File Types Supported for Select AI RAG Vector Index Creation	Select AI Retrieval Augmented Generation (RAG) supports multiple file formats, including PDF, DOCX, JSON, XML, and HTML, among others. For more information, see Select AI with Retrieval Augmented Generation (RAG) and, for the complete list of supported file formats, Build your Vector Store.

Parent topic: What’s New for Oracle Autonomous Database Serverless

March 2025

Feature	Description
Elastic Pool with Any Workload Leader	When you create an elastic pool, the workload type of the leader is not restricted. The workload type of an elastic pool leader can be one of: Transaction Processing, Data Warehouse, JSON, or APEX. See About Elastic Pools and About Autonomous Database Workload Types for more information.
SQL Parse Error Warning Event: `SQLParseErrorWarning`	If a SQL parse error occurs several times in a certain time window, for example 200 times or more in an hour, this generates `"Too Many Parse Errors"` entries in the alert log. The `SQLParseErrorWarning` event is triggered when there is a new entry in the alert log, indicating that a SQL statement has too many parse errors. For example, when an invalid query runs 200 times in less than an hour, this triggers the `SQLParseErrorWarning` event. See Information Events on Autonomous Database for more information.
Create Refreshable Clone with 19c Source and 23ai Refreshable Clone	When you create a refreshable clone from an Oracle Database 19c source, you can select Oracle Database 23ai for the refreshable clone. You can use the refreshable clone to test your database on Oracle Database 23ai. After your testing is complete, you can disconnect the refreshable clone to upgrade to Oracle Database 23ai. See Upgrade an Autonomous Database Instance to Oracle Database 23ai and Create a Refreshable Clone for an Autonomous Database Instance for more information.
Offload Queries from Elastic Pool Leader to Member Refreshable Clones: Includes Queries from PL/SQL and from Oracle Scheduler Jobs	When you have heavy read workloads, where a relatively small amount of data is scanned by multiple queries, including from PL/SQL or from Oracle Scheduler Jobs, you can offload queries (reads) from an elastic pool leader to Refreshable Clones. See Offload Queries from Elastic Pool Leader to Member Refreshable Clones for more information.
Connections Dropped Metrics	The `ConnectionsDroppedByClient` metric shows the sum of the number of connections dropped by the client. The `ConnectionsDroppedByServer` metric shows the sum of the number of connections dropped by the server. See Available Metrics: oci_autonomous_database for more information.
Oracle APEX 24.2	Autonomous Database uses Oracle APEX Release 24.2. See Create Applications with Oracle APEX in Autonomous Database for more information.
Data Studio Enhancements	The Data Studio is updated with new features and changes: The Table AI Assist Tool: The Data Studio suite of tools now features an AI-powered Table Assist Tool to effortlessly enhance your local data tables with the power of AI. You can use natural language prompts to augment and correct your information with this tool. See The Table AI Assist Tool for more information. Catalog Tool Enhancements: The new multi-catalog capabilities of the Autonomous Database catalog mean you can add catalogs to enable data search, discovery and use across a wide range of external systems, including other Autonomous Databases, on-premises databases, cloud storage systems, shared data, and other data catalogs such as Oracle Cloud Infrastructure Data Catalog and AWS Glue. See The Catalog Tool for more information. The Data Marketplace Tool: Using Data Studio Marketplace, Autonomous Database users can easily collaborate by sharing data with other users of the same or another Autonomous Database. Data providers publish data sets in just minutes, no coding required. Data consumers can browse available data sets, and query live data or copy it to their own Autonomous Database Instance. See The Data Marketplace Tool for more information. Data Load Livefeed Enhancements: With the new Basic mode, creating a Livefeed is as simple as selecting a folder from a drop-down menu. New ability to use Merge semantics in livefeeds. The Livefeed feature automatically updates existing data in your tables with new information, eliminating redundant rows and ensuring data integrity. With newly introduced Show code icon, you can view the underlying SQL code behind your livefeeds and even run the code manually in the SQL worksheet. See Feeding Data for more information.
Documentation Addition: Migrate Existing SQL Execution Plans to Autonomous Database	You can use Real-time SQL Plan Management when you migrate from a source Oracle Database to Autonomous Database. See Migrate Existing SQL Execution Plans to Autonomous Database to Mitigate the Risk of SQL Performance Regressions for more information.

Parent topic: What’s New for Oracle Autonomous Database Serverless

February 2025

Feature	Description
Offload Queries from Elastic Pool Leader to Member Refreshable Clones	When you have heavy read workloads, where a relatively small amount of data is scanned by multiple queries, you can offload queries (reads) from an elastic pool leader to Refreshable Clones. See Offload Queries from Elastic Pool Leader to Member Refreshable Clones for more information.
Create External Tables Over PAR URLs	You can create an external table over Pre-Authenticated Request (PAR) URLs. See Query Data that Uses Table Hyperlink Source for more information.
Update Cloud Links Registrations	With Cloud Links, a data owner registers a table or view for remote access for a selected audience as defined by the data owner, and the data is then accessible to those with access granted at registration time. After you register a data set you can update data set attributes. See Update Data Set Registration Attributes for more information.
Database User Password Warning Event: `DatabaseUserPasswordWarning`	The database user password warning event provides a message that the Autonomous Database user password is expiring within the specified grace period (default 30 days) or has already expired. See Critical Events on Autonomous Database for more information.

Parent topic: What’s New for Oracle Autonomous Database Serverless

January 2025

Feature	Description
Documentation Addition: Create and Use Custom Metrics	You can create custom metrics to collect, publish, and analyze metrics. A custom metric collects specified data from your Autonomous Database instance and publishes the data to the OCI Monitoring service using a REST API provided by the OCI SDK. See Use Custom Metrics on Autonomous Database for more information.
Disconnect a Peer Database	You can disconnect an Autonomous Data Guard standby or a Backup-Based Disaster Recovery peer. When you disconnect a cross-region peer, the database is permanently disassociated from the Primary database. This converts the database from a peer database to a standalone database. See Disconnect a Peer Database for more information.
Change Patch Level After Provisioning	When you provision or clone an Autonomous Database instance you can select a patch level to apply to upcoming patches. You can also change the patch level after an Autonomous Database instance is provisioned. There are two patch level options: Regular and Early. See Set the Patch Level for more information.
Service Name Mapping for Cloud Links Consumers	You can set the service name mapping to use when Cloud Links consumers access data from a data set owner. See Set Database Service Name Mapping for Cloud Links Consumers for more information.
External Tables with ORC, Parquet, and Avro Provide an Option to Set Size of String Columns	When you access ORC, Parquet, or Avro data stored in object store using external tables, you can either automatically or explicitly set the maximum size of text columns. By default the text column size is set based on the value of `MAX_STRING_SIZE`. See Query External Data with ORC, Parquet, or Avro Source Files and Explicitly Set the Text Column Size for more information.
Invoke Google Cloud Platform Cloud Run Functions	You can invoke Google Cloud Platform Cloud Run functions as SQL functions. See Steps to Invoke Google Cloud Run Functions as SQL Functions for more information.

Parent topic: What’s New for Oracle Autonomous Database Serverless