Oracle AI Database 26ai — New Features Quick Reference

Over 300 new features focused on AI and developer productivity • Generated by oramad on 15-Feb-2026

11 Major Categories 300+ Features AI & Vector Search JSON Duality SQL/PGQ Graphs Post-Quantum Crypto
AI Vector Search 40 features
General & Vector Data Type
Vector Data TypeBuilt-in VECTOR data type enabling AI-powered similarity searches directly within the database.
Sparse VectorsSupport for high-dimensional vectors with few non-zero values, ideal for keyword-sensitive search models like SPLADE and BM25.
BINARY Vector Dimension FormatSingle-bit-per-dimension format offering 32x less storage and up to 40x faster distance computations.
Arithmetic & Aggregate Operations on VectorsApply +, -, * arithmetic and SUM/AVG aggregates directly on vector columns.
AI Vector Search: SQL ExecutionSQL execution support for vector indexes, vector functions, and row-level restrictions on partitions.
AI Vector Search: OptimizerOptimizer can now use vector indexes instead of full table scans for efficient vector queries.
AI Vector Search: PL/SQLNew vector type in PL/SQL with operations for similarity searches within PL/SQL blocks.
PL/SQL BINARY Vector SupportPL/SQL supports BINARY dimension format for vectors, matching SQL capabilities.
Sparse Vector Support in PL/SQLNative sparse vector creation and usage directly from within PL/SQL.
Dimension-Wise Arithmetic in PL/SQLVector arithmetic (+, -, *) now available natively in PL/SQL without calling out to SQL.
Support of Vector in JSON Type (OSON)Vector data type embedded in JSON scalar types for interoperability between SQL and JSON.
User-Defined Vector Distance FunctionsCreate custom distance metrics using JavaScript for domain-specific similarity searches.
Vector Data Type in External TablesAccess and search vector data stored outside the database via external tables.
JDBC Support for Vector Data TypeFull JDBC driver support for the VECTOR data type including metadata, conversions, and binding.
OCI Support for Vector TypeOracle Call Interface supports the VECTOR data type for OCI-based applications.
Client Support for SPARSE VectorsJDBC and ODP.NET client drivers now support sparse vector types.
JDBC Support for SPARSE Binary VectorsSparseBooleanArray interface added for sparse binary vector data in Java applications.
Sharding Support for AI Vector SearchVector tables auto-distributed across shards with parallelized similarity searches.
Set COMPATIBLE to 23.6.0New vector search features require COMPATIBLE parameter set to 23.6.0.
Vector Indexes
Hybrid Vector Index (HVI)Combined full-text search and semantic vector search in a single index for higher quality results.
Vector Indexes (IVF)Specialized indexing for efficient approximate similarity searches on high-dimensional data.
Partition-Local Neighbor Partition Vector IndexLOCAL indexing for partitioned tables enabling partition-pruned vector searches.
Persistent Neighbor Graph Vector IndexesDisk checkpoint for HNSW graphs to speed up index reload after database restart.
Transactional HNSW IndexesDMLs now allowed on tables with HNSW indexes with transactionally consistent search results.
Distributed HNSW Indexes with RACHNSW indexes scale across all RAC instances' total available memory.
Duplicated HNSW on RACFull HNSW index duplication across all RAC cluster instances for ultra-fast searches.
Scalar Quantized HNSW IndexesCompress HNSW indexes using scalar quantization to reduce storage footprint.
Auto IVF Vector Index ReorganizationAutomatic rebuild of IVF indexes when performance degradation is detected.
IVF Index Online ReorganizationReorganize IVF indexes while they remain available for DMLs and queries.
Online Index RebuildCreate or rebuild vector indexes while the base table stays open for updates.
Included Columns in Vector IndexesStore non-vector columns in IVF/HNSW indexes to eliminate base table lookups.
Hybrid Vector Index for JSONHybrid vector index extended to support JSON columns for text + semantic search.
Partition Maintenance with Global Vector IndexesAdd, drop, merge, split partitions on tables with global IVF and HNSW indexes.
Additional Predicate Support in Hybrid SearchWHERE clause predicates on non-indexed columns now supported in hybrid vector search.
Distance Functions & ONNX
Jaccard Distance MetricNew Jaccard distance metric for measuring similarity between binary vectors.
PL/SQL JACCARD Distance SupportJACCARD distance metric available in PL/SQL VECTOR_DISTANCE operator.
ONNX-Format Models as DB ObjectsImport ONNX models as first-class MINING MODEL objects for in-database inference.
Enable Larger ONNX ModelsSupport for ONNX models larger than 1GB using external initializers.
Image Transformer Models via ONNX RuntimeImport and use image transformer models for in-database image vectorization.
Vector Utilities & Memory
Vector Utility API (VECTOR_CHUNKS)SQL function to chunk text into pieces for embedding generation.
Chainable Utility FunctionsPipeline text-to-embedding functions together for streamlined vector index creation.
Select AI with RAG & Synthetic DataNatural language to SQL with retrieval augmented generation and synthetic data generation.
Hybrid Vector Search API (DBMS_HYBRID_VECTOR)JSON_EXISTS and JSON_TEXTCONTAINS support in hybrid search queries.
Elastic Vector Memory ManagementVector memory pool auto-resizes to accommodate HNSW index changes dynamically.
Vector Memory Pool Auto ManagementAutonomous DB automatically manages vector pool sizing for HNSW indexes.
Application Development 95+ features
JSON & JSON-Relational Duality
JSON-Relational Duality ViewsFully updatable JSON views over normalized relational data, combining document and relational benefits.
JSON Schema ValidationValidate JSON data structure and types using IS JSON condition and PL/SQL utility.
JSON CollectionsNative document-store-compatible JSON tables/views integrated with SQL.
JSON Collection ViewsRead-only views exposing JSON objects, conceptually close to duality views with fewer restrictions.
GoldenGate Replication of Duality ViewsReplicate duality view data as JSON documents to Oracle and non-Oracle databases.
JSON Replication (Partial JSON Logging)Log only modified portions of JSON for efficient replication, reducing network overhead.
JSON-to-Duality MigratorPL/SQL tools to migrate document collections from document databases to duality views.
Enhancement to JSON_TRANSFORMExtended with RHS path expressions, nested paths, arithmetic operations, and array sorting.
GraphQL Syntax for Duality ViewsEnhanced GraphQL directives, comments, QBE predicates for creating duality views.
JSON Type ModifiersConstrain JSON columns to objects, arrays, or scalars with size and element limits.
Comparing and Sorting JSON TypesUse JSON data type directly in WHERE, ORDER BY, and GROUP BY clauses.
JSON_EXPRESSION_CHECK ParameterDetect incorrect JSON path expressions at compile time instead of runtime.
Simple Dot-Notation for JSON IndexesCreate indexes on JSON data using simple dot notation for improved query performance.
WHERE Clauses in Duality ViewsLimit rows in duality views with WHERE clauses for fine-grained document control.
Precheckable Constraints using JSON SchemaApplications can precheck data validity using JSON schemas before sending to the database.
ORDERED in JSON_SERIALIZESort key-value pairs alphabetically in JSON_SERIALIZE output for easy comparison.
DBMS_AQ Support for JSON ArraysUse JSON array payloads for Advanced Queuing bulk message passing.
SODA EnhancementsMerge, patch, embedded keys, sampling, flashback, and more for SODA API.
SQL Enhancements
BOOLEAN Data TypeISO SQL standard-compliant BOOLEAN type for storing TRUE/FALSE values.
Direct Joins for UPDATE and DELETEJoin target table to other tables using FROM clause in UPDATE/DELETE statements.
IF [NOT] EXISTS SyntaxDDL creation/modification/deletion supports conditional existence checks.
SELECT Without FROM ClauseRun expression-only SELECT queries without requiring a FROM clause.
GROUP BY Column Alias or PositionUse column aliases or positions in GROUP BY, ROLLUP, CUBE, and GROUPING SETS.
GROUP BY ALLAutomatically group by all non-aggregated columns without listing them.
Table Value Constructor (VALUES)Materialize rows on the fly in SELECT, INSERT, and MERGE using VALUES clause.
QUALIFY ClauseFilter results of analytic functions without nesting queries.
Non-Positional INSERT (SET & BY NAME)Self-documenting INSERT SET clause and BY NAME matching for subquery inserts.
DEFAULT ON NULL for UPDATEColumns auto-set to default value when an UPDATE tries to set NULL.
Aggregation Filters (FILTER Clause)Apply conditions directly within aggregate functions per SQL standard.
Aggregation over INTERVAL TypesPass INTERVAL data types to SUM and AVG aggregate/analytic functions.
Time Bucketing (TIME_BUCKET)Native SQL operator for fixed-time interval bucketing of time series data.
Calendar FunctionsExtract year, quarter, month, week, day across calendar, fiscal, and retail calendars.
DATEDIFF / TIMESTAMPDIFFStandard date/time difference calculations for easier migration from other databases.
UUID() FunctionGenerate RFC 9562-compliant version 4 UUIDs for unpredictable unique identifiers.
UPDATE RETURN Clause EnhancementsRETURNING INTO clause reports old and new values for INSERT/UPDATE/DELETE/MERGE.
Extended CASE ControlsDangling predicates and multiple choices in WHEN clauses per SQL:2003 standard.
Schema AnnotationsStore and retrieve application metadata as name-value pairs on database objects.
Data Use Case DomainsCentralized declaration of column intent (email, URL, currency) with constraints.
Unicode 15.0 SupportNLS data files updated to match Unicode Standard version 15.0.
Automatic PL/SQL to SQL TranspilerPL/SQL functions in SQL automatically converted to SQL expressions for speed.
Materialized View Rewrite with Bind VariablesQuery rewrite now supports SQL with bind variables via bind peeking.
Property Graph
Native Property Graphs in SQLBuilt-in property graph data structures and queries directly in Oracle AI Database.
ISO SQL/PGQ Standard SupportAmong the first to support ISO SQL standard for property graph queries.
Graph from JSON CollectionsExecute SQL/PGQ queries on graphs represented as JSON columns.
Graph from Database ViewsCreate property graphs from views, database links, external tables, and more.
RDF Graph AnalyticsUse PageRank, Community Detection, and other graph algorithms with RDF graphs.
Microservices & Messaging
Kafka APIs for TxEventQRun Kafka applications directly against the database using Transactional Event Queues.
TxEventQ PropagationPropagate events/messages between databases as a reliable and secure Event Mesh.
Saga APIs (Oracle Saga Framework)Framework for maintaining atomic data consistency across microservices transactions.
Python & REST Drivers for TxEventQTxEventQ now accessible via Python and Kafka-compatible REST APIs.
Prometheus/Grafana for OracleDatabase metrics dashboards for developers in Kubernetes/Docker environments.
ODP.NET: AQ and TxEventQ.NET drivers support Advanced Queuing and Transactional Event Queues APIs.
JavaScript & Java
MLE JavaScript Modules & EnvironmentsPersist and manage JavaScript code as schema objects in the database.
JavaScript Module Calls from SQL/PL/SQLInvoke JavaScript functions stored in modules from SQL and PL/SQL via call specs.
JavaScript SODA APIClient-side and server-side SODA document access for JavaScript developers.
Foreign Function Interface (JS to PL/SQL)Use native JavaScript constructs to call PL/SQL code units directly.
Restricted Execution Contexts (PURE)Limit JavaScript to pure computation without database interaction for safety.
JavaScript Post-Execution DebuggingCollect runtime state during execution for post-mortem debugging of JS code.
Java in DB: JDK 11 + ModulesRe-architected Oracle JVM supporting JDK 11 capabilities and the Java module system.
Java Web Services Callout EnhancementMore efficient web services callout from Java, PL/SQL, and SQL in the database.
Application Connectivity & Drivers
Reset Database Session StateAutomatically clean session state at request end to prevent data leaks.
Implicit Connection Pooling (DRCP)Auto-assign DRCP servers to/from connections without explicit app pooling.
Multiple Named DRCP PoolsCreate multiple named pools for finer control over DRCP usage.
OCI Pipelined OperationsSubmit multiple database operations without waiting for server responses.
ODP.NET: Application Continuity.NET drivers support transparent recovery of in-flight sessions during outages.
ODP.NET: PipeliningQueue database requests transparently while awaiting responses for better throughput.
ODP.NET: OpenTelemetryInstrumented APIs for distributed tracing using the OpenTelemetry framework.
JDBC Pipelined Database OperationsAsynchronously submit multiple SQL requests without waiting for preceding calls.
JDBC OAuth 2.0 (OCI IAM & Azure AD)Simplified Java authentication using OAuth 2.0 in multi-cloud environments.
Java Support for True CacheJDBC transparently routes read-only workloads to True Cache instances.
DBMS_DEVELOPER PackageEfficient JSON-format metadata retrieval for database objects replacing XML-based methods.
GraphQL Table Function & Schema APIQuery Oracle Database using GraphQL language via SQL table function.
SQLPlus Enhancements (PING, OERR, CONFIG)New commands for connectivity testing, error lookup, and configuration management.
Data Analytics 35+ features
General Analytics
Automatic Data ClusteringTransparently clusters data based on workload for better zone map/storage index pruning.
Data Quality OperatorsPHONIC_ENCODE and FUZZY_MATCH for pronunciation-based and fuzzy string matching.
Oracle SQL Access to KafkaQuery or load Apache Kafka/OCI streaming data via SQL using DBMS_KAFKA.
Hybrid Partitioned Tables (Interval/Auto-List)Hybrid partitioned tables now support interval and automatic list partitioning.
Text Indexes with Automatic MaintenanceBackground auto-sync for text indexes eliminates manual SYNC_INDEX operations.
Enhanced Automatic IndexingAccounts for DML costs, supports range predicates and function-based indexes.
Enhanced Automatic Materialized ViewsAuto-partitioning and improved cost model for automatic MV selection and refresh.
Enhanced Automatic SQL Plan ManagementFaster detection and repair of SQL plan regressions at parse-time.
Concurrent MV Refresh (on-commit)Multiple sessions can refresh the same on-commit materialized views simultaneously.
MV Support for ANSI JoinsFull rewrite capabilities for materialized views with ANSI join syntax.
Semi-Join Materialized ViewsReplace large unified dimension tables with join-specific MVs for faster analytics.
Ubiquitous Search (DBMS_SEARCH)Index multiple schema objects in a single index for cross-object full-text search.
Transportable Binary XMLSelf-contained XMLType storage supporting sharding, search index, and Exadata pushdown.
In-Memory
Automatic In-Memory EnhancementsAuto-enable/disable join groups, hashed dictionaries, and optimized arithmetic.
Automatic In-Memory Sizing (ADB)In-Memory column store grows/shrinks dynamically based on workload for Autonomous DB.
Vectorized Query ProcessingSIMD-exploiting multi-level hash joins and group-by aggregations for faster analytics.
In-Memory RAC-Level Global DictionarySynchronized common dictionaries across RAC nodes for improved join performance.
In-Memory Optimized DatesExtract and cache DATE components in IM column store for faster date-based queries.
Selective In-Memory ColumnsALL sub-clause to easily enable/disable all columns for in-memory.
Machine Learning
Automated Time Series Model SearchAuto-select Exponential Smoothing model type and hyperparameters for forecasting.
XGBoost: Constraints & Survival AnalysisXGBoost supports survival analysis (AFT model) and feature interaction constraints.
GLM Link Functions (probit, cloglog, cauchit)Additional link functions for logistic regression expanding model quality options.
Multiple Time SeriesConstruct time series regression models with multivariate inputs and indicator data.
Outlier Detection using EM ClusteringDistribution-based anomaly detection via Expectation Maximization algorithm.
ESA Dense Projection with EmbeddingsExplicit Semantic Analysis outputs doc2vec representations for text analytics.
OML4Py & OML4R EnhancementsAdditional algorithms (NMF, ESM, XGBoost, Random Forest, Neural Network) for Python and R.
In-Database Vector PredictorsUse VECTOR columns as predictors alongside structured data in ML algorithms.
Partitioned Model PerformanceImproved performance for up to 32K partitions in partitioned models.
Spatial
Spatial: 3D Models & AnalyticsPoint cloud change detection via SQL/PL/SQL for landscape and infrastructure monitoring.
Spatial: REST APIs for GeoRasterREST APIs for querying and manipulating satellite imagery and raster data.
Spatial Vector Tiles (PL/SQL API)Generate vector tiles from geometries for efficient web map rendering.
Data Warehousing / Big Data 7 features
Enhanced Partitioning MetadataPartition boundary info available in JSON and CLOB format for programmatic use.
Extended Language Support in Oracle TextUp to 48 languages supported with on-demand language file downloads.
External Table Partition Values in File PathDerive partition values from directory/file names for Hive-style external tables.
Logical Partition Change TrackingFine-grained MV staleness tracking at logical partition level without MV logs.
Staging TablesHeap tables pre-optimized for fast data ingestion and volatile data handling.
JSON Materialized Views (Enhanced)Faster refresh and broader query rewrite for JSON table materialized views.
Heat Map Retention TimeControl how long Heat Map access/modification tracking data is retained.
Cloud Migration 2 features
Classic Queue to TxEventQ MigrationOnline tool to migrate from AQ Classic Queues to high-performance TxEventQ.
Cloud Premigration Advisor Tool (CPAT)Proactively identifies and recommends fixes for migration incompatibilities.
Cloud Operations 15 features
Manageability
Hybrid Read-Only Mode for PDBsPDB operates read-write for common users, read-only for local users during maintenance.
Real-Time SQL Monitoring EnhancementsPer-PDB/CDB monitoring, AWR export, and SQL History Reporting integration.
Control PDB Open OrderDefine startup priority for PDBs so critical databases open first.
Inter-Instance Resource ManagementFine-grained CPU control with limits and guarantees across CDBs, PDBs, and processes.
Read-Only Users and SessionsDisable/re-enable read-write capabilities for any user or session without revoking privileges.
Continuous Availability & General
App Continuity with Database TemplatesCheckpoint and restore session state using templates for broader AC protection.
Smart Connection RebalanceAutomatically reshuffle RAC connections based on real-time performance monitoring.
Smooth RAC ReconfigurationReduced brownout time when nodes join/leave an Oracle RAC cluster.
Adaptive Result Cache ExclusionAuto-blocklist objects where result caching is not beneficial.
Cloud Developer PackagesBuilt-in DBMS_CLOUD packages for cloud storage, pipelines, notifications, and AI integration.
Unified MemorySingle MEMORY_SIZE parameter replaces SGA_TARGET and PGA_AGGREGATE_LIMIT.
Read-Only Tablespace on Object StorageMove read-only tablespaces to Oracle Object Storage for cost-effective tiering.
Azure Blob Storage BackupRMAN backup and recovery directly to/from Microsoft Azure Blob Storage.
Auto SQL Compile-Time Error RepairAutomatically detect and repair severe compile-time SQL exceptions.
High Availability 18 features
Data Guard
Data Guard Redo Decryption (Hybrid DR)Decrypt redo in hybrid cloud DR where cloud DB is TDE-encrypted and on-prem is not.
Per-PDB Data Guard EnhancementsIndependent switchover/failover per PDB with simplified setup and Real-Time Query.
Configuration & Member TaggingUser-defined key-value tags for Data Guard members for logical grouping.
JSON Output for DGMGRLData Guard command-line output available in JSON for automation integration.
Restrict Switchover/Failover CandidatesPrimaryDatabaseCandidates property controls which databases can become primary.
Optimized Fast-Start Failover Lag DetectionNew lag type and grace time properties for improved Maximum Performance mode.
General HA & RAC
Flashback Time Travel EnhancementsAutomatically track and archive transactional changes with schema evolution history.
Database Native Transaction GuardEnhanced Transaction Guard without extra redo generation or performance overhead.
Oracle RAC Two-Stage Rolling UpdatesApply non-rolling fixes in rolling fashion, reducing downtime for RAC patches.
General Purpose Cluster ConfigurationSimplified cluster deployment with minimal network/storage configuration.
Local Rolling Database MaintenanceRolling patching locally on one RAC node without affecting other cluster nodes.
Single-Server Rolling MaintenanceRolling patching on a single server hosting RAC or RAC One Node databases.
Oracle DBCA: Standard Edition HAQuickly create Standard Edition HA databases with automatic failover via DBCA.
Application Continuity for DBMS_ROLLINGContinuous availability during rolling upgrades using Application Continuity.
Security 30+ features
SQL Firewall & Encryption
Oracle SQL FirewallReal-time protection blocking unauthorized SQL and SQL injection attacks, built into the database.
TLS 1.3 SupportLatest TLS protocol with improved session setup performance and stronger ciphers.
Post-Quantum Cryptography (ML-KEM, ML-DSA)Quantum-resistant cryptographic algorithms for future-proof data protection.
AES-XTS Encryption for TDE TablespaceXTS mode for tablespace encryption with improved security and parallel performance.
TDE Default Changed to AES256Default encryption algorithm upgraded to AES256 for both column and tablespace encryption.
Simplified TLS ConfigurationStreamlined parameters and client wallets not required for well-known CA certs.
Improved Local Auto-Login WalletsMore tightly bound to host, more secure, no root access required.
Changes to DBMS_CRYPTOAdded AES-XTS mode, SHA-3, and SM2/3/4 to cryptographic package.
Authentication & Authorization
Password Length Increased to 1024 BytesDatabase and client passwords now support up to 1024 bytes for stronger authentication.
Multi-Factor AuthenticationMFA via push notifications (Cisco Duo/Oracle Mobile Auth) or PKI certificates.
Microsoft Azure AD IntegrationSingle sign-on to Oracle databases using Azure AD OAuth2 tokens.
Schema PrivilegesGrant privileges at schema level instead of broad system-wide ANY privileges.
DB_DEVELOPER_ROLEPre-built role with all necessary privileges for application development.
Updated Kerberos Library (v1.21.2)MIT Kerberos 1.21.2 with cross-domain support for multi-domain environments.
RADIUS Enhancements (RFC 6613/6614)TCP over TLS by default for RADIUS MFA authentication.
OCI IAM IntegrationLog in to OCI DBaaS databases using IAM password or token-based authentication.
Audit & Other
Column-Level AuditingCreate unified audit policies for individual columns in tables and views.
Database Vault Audit ControlsControl AUDIT_ADMIN/AUDIT_VIEWER role grants via Database Vault APIs.
Data Dictionary Protection ExtendedNon-SYS Oracle schemas get data dictionary protection with separation of duties.
Oracle Data Redaction EnhancementsOptimized capabilities and removed previous limitations for data redaction.
IP Rate Limiting in CMANLimit new connections per IP to protect against denial-of-service attacks.
Digital Signature for Installation ArchivesOracle-signed installation files ensure integrity before deployment.
Secure Distributed Transaction RecoveryAdditional security for the RECO background process in distributed transactions.
OLTP & Core Database 45+ features
Availability & Sharding
True CacheIn-memory, consistent, diskless cache for read-only workloads with automatic management.
Raft Replication for Globally Distributed DBBuilt-in consensus-based replication with sub-second failover, no GoldenGate needed.
Directory-Based ShardingDynamic data placement with key location stored in a directory for flexible scaling.
Priority TransactionsAuto-abort lower priority transactions blocking higher priority ones past timeout.
Lock-Free ReservationReserve column values without row locks for improved concurrency.
Sessionless TransactionsSuspend/resume transactions across sessions without external transaction managers.
Automatic Data Move on Sharding Key UpdateData auto-moves to correct shard/partition when sharding key value changes.
Automatic Transaction QuarantineQuarantine problematic transactions while keeping the database open and operational.
Immutable Backups Using RMANRMAN compatible with immutable OCI Object Storage for ransomware protection.
Parallel Cross-Shard DMLCross-shard updates and inserts execute in parallel across multiple shards.
Blockchain & Immutable Tables
Blockchain Table User ChainsUser-defined chains for isolated verification based on custom column values.
Blockchain Table Row VersionsMaintain multiple historical versions of rows within blockchain tables.
Blockchain Table Log HistoryTrack changes to regular tables in cryptographically secure blockchain audit log.
Add/Drop Columns in Blockchain TablesEvolve blockchain/immutable table schemas while preserving crypto-hash chain integrity.
Blockchain Table CountersignatureDatabase countersigns rows for non-repudiation with external storage options.
Blockchain Table Delegate SignerAllow alternate users to sign blockchain table rows on behalf of the primary user.
Database Architecture
Wide Tables (4096 Columns)Maximum columns per table increased from 1000 to 4096 for ML/IoT workloads.
Partition by ExpressionDefine partition keys using expressions directly, eliminating virtual column overhead.
Assertions for Data IntegrityDeclarative multi-table business rules as database-level assertions.
Unrestricted Direct LoadsQuery and DML allowed on same table after direct load before commit.
Unrestricted Bulk TransactionsMultiple parallel DMLs and queries on same object within a single transaction.
System Timezone Autonomy for PDBsEach PDB controls its own SYSDATE/SYSTIMESTAMP timezone independently.
Memory Speed ColumnarNext-gen HCC with in-memory format, eliminating columnar format mismatch across tiers.
HCC Compression ImprovementsFaster compression/decompression speeds and better ratios for Hybrid Columnar Compression.
Manageability & Performance
Bigfile Tablespace ShrinkReliably shrink bigfile tablespaces to reclaim unused space.
Smallfile Tablespace ShrinkReliably shrink smallfile tablespaces to close to actual used space.
BIGFILE Default for SYS TablespacesSYSAUX, SYSTEM, and USER tablespaces default to bigfile, reducing datafile count.
Automatic SecureFiles ShrinkBackground auto-shrink of SecureFile LOB segments without manual intervention.
Automatic Storage CompressionDirect load into uncompressed format, then gradually move to HCC in background.
Fast Ingest EnhancementsMemoptimize for Write now supports partitioning, compressed tables, and direct In-Memory population.
Ordered Sequence Optimizations in RACBetter performance for ordered sequences in RAC without manual tuning.
Materialized Expression ColumnsPersist virtual columns on disk to avoid repeated computation.
SQL*Loader Auto Parallel Direct LoadAutomatic parallel loading without splitting data files or starting multiple clients.
Centralized Configuration ProvidersPull connection descriptors and config from Azure or OCI cloud stores.
Enhanced Query History TrackingTrack complete query history including sub-5-second non-parallel queries.
RMAN Enhancements (AES256 Default)Encrypted backups default to AES256; includes cross-platform migration improvements.
Diagnosability 15 features
Database Observability (KSTRC)Distributed tracing with OpenTelemetry integration for end-to-end issue debugging.
Telemetry StreamingTurnkey metrics streaming with REST/PL/SQL ingestion and PromQL querying.
SQL Diagnostic Report (DBMS_SQLDIAG)Deep-level HTML diagnostic report for SQL statements with plan history and statistics.
Network Monitor (NMON)Always-on network monitoring with health scores at system, interface, IP, and flow levels.
Oracle Update AdvisorSoftware update recommendation framework for keeping software at recommended versions.
Cluster Health Monitor ImprovementsListen for critical events, recommend corrective actions, and self-heal autonomously.
Enhanced Cluster Health AdvisorSupport for 4K PDBs (up from 256) with improved problem detection accuracy.
DBMS_DICTIONARY_CHECKLightweight PL/SQL package to identify and remediate data dictionary inconsistencies.
In-Memory Advisor (Built-in)Built-in eligibility test and enhanced analysis to identify Database In-Memory candidates.
Add Verified SQL Plan BaselineAutomatically find and baseline the best execution plan from cache, AWR, or tuning set.
CMAN Diagnostics EnhancementsMonitor service registration stats and diagnose registration issues in CMAN logs.
OCI Client-Side Tracing APIsDynamic OCI diagnostic tracing without configuration file changes.
HA Diagnosability (DBMS_SCHEDULER)In-memory tracing for scheduler jobs during forced shutdowns and patching.
Extent-Based ASM ScrubbingScrub specific extents instead of entire files for faster turnaround and less impact.
Installation, Upgrade & Patching 14 features
Installation
DB Installer Command-Line SupportSpecify install commands and parameters via CLI in addition to GUI.
Grid Infrastructure Installer ImprovementsGold image management and out-of-place patching with reduced inventory metadata.
Oracle FPP Local Mode (No Java Container)Fleet Patching without GIMR or Java container for simpler, faster patching.
Store/Transfer Gold Images as ZIPSave storage and bandwidth by storing/transferring FPP gold images as ZIP files.
Archiving/Unarchiving Gold ImagesArchive unused FPP gold images to external storage to free repository space.
Upgrade
AutoUpgrade Release Update (RU) UpgradesOut-of-place Oracle home patching with AutoUpgrade for seamless RU application.
AutoUpgrade Auto-ParallelismAuto-evaluates system resources to determine simultaneous upgrade job count.
AutoUpgrade TDE Keystore SupportAutomated keystore management for TDE-enabled database upgrades.
AutoUpgrade Unplug-Plugin to Different SystemsUnplug PDB from one server, plug into another, and upgrade in a single operation.
REST APIs for AutoUpgradeORDS and OCI REST APIs for secure remote AutoUpgrade operations.
Patching
Gold Image Out-of-Place Patching (DBCA)DBCA applies RUs out-of-place using gold images for Grid Infrastructure and DB homes.