This release (Release 2026.03.13) introduces a series of new features, enhancements, and bug fixes.

• Domestic Regions

  • Alibaba Cloud (Shanghai)
  • Tencent Cloud (Shanghai/Beijing/Guangzhou)
  • AWS (Beijing)

• International Regions

  • Alibaba Cloud (Singapore)
  • AWS (Singapore)

AI Foundation Capabilities

• Unstructured Data Processing: Officially launched the Stream on Volume feature, supporting incremental change capture for unstructured files within Volumes. Users can create Streams on Volumes to automatically track newly added file changes, enabling efficient incremental processing of continuously written unstructured data such as images, PDFs, audio/video, and logs, without the need for full scans each time.
• AI Functions: Officially launched two built-in AI functions, allowing users to invoke large language model capabilities directly within SQL, enabling intelligent data processing without leaving the data platform. See: AI_COMPLETE and AI_EMBEDDING

New Features

Dynamic Table Inverted Index Support

• The DDL syntax for Dynamic Tables now supports declaring inverted indexes at creation time. Users can directly specify Inverted Indexes in the CREATE DYNAMIC TABLE statement, enabling efficient keyword search capabilities on text columns within dynamic tables without the need for additional index maintenance after table creation. This capability is particularly suitable for scenarios requiring full-text search or hybrid queries on incrementally materialized results.

Resource Management

• Added Compute Pool (VC Resource Pool) Feature [OP Version Only]: The Compute Pool feature can be used to isolate privately deployed compute cluster (VC) resource pools, ensuring that each resource pool has a minimum available compute resource that cannot be preempted by other resource pools. A single service instance can create multiple compute pools (VC resource pools), and each compute pool can contain multiple VCs (virtual clusters).

SQL Feature Enhancements

SQL Commands

• Show Partitions Enhancement: show partitions now supports descending order. Reference Link
• Kafka PIPE Supports Consumer Offset Modification: Use ALTER PIPE ... SET RESET_KAFKA_GROUP_OFFSETS='latest | earliest' to modify consumer offsets. The PIPE must be stopped before executing the modification command.
• INSERT Enhancement: The INSERT statement now supports UPDATE ON PK CONFLICT.
• Dynamic Table Index Enhancement: Supports setting INVERTED INDEX in CREATE OR REPLACE DYNAMIC TABLE.

Security

• Added AWS Private Link Support: Used for creating and connecting to data sources within the AWS environment. This avoids data transmission over the public Internet, improving data security.

System Behavior Changes

• When creating a new workspace, a sync-type cluster named 'DEFAULT_INTEGRATION' is now created by default to facilitate data integration tasks within that workspace. This cluster is stopped by default and will not incur compute resource consumption or charges when creating a new workspace. Existing workspaces are unaffected.