In this release (Release 2025.07.15), we are introducing a series of new features, enhancements, and fixes. These updates will be rolled out gradually to the following regions, with completion expected within one to two weeks from the release date. Specific timing depends on your region.

Global Regions

• Alibaba Cloud (Singapore)
• AWS (Singapore)

China Regions

• Alibaba Cloud (Shanghai)
• Tencent Cloud (Shanghai / Beijing/ Guangzhou)
• Amazon Web Services (Beijing)

New Features

Spark On Lakehouse Productization [Private Preview]

• Volume File Referencing Support: When submitting Spark jobs, you can now use Volume addresses to reference program files and dependency files.
• Storage Integration: Supports Spark reading from and writing to Lakehouse storage, compatible with Lakehouse Catalog and internal table read/write operations.
• Job History Integration: Spark jobs are now visible and displayed within the Studio Job History.
• Lakehouse Studio Integration: Added single-tenant services, job monitoring, and management APIs to strengthen end-to-end control capabilities.

Security & Permission Enhancements

• Kerberos Authentication: HDFS connections now support Kerberos authentication, enhancing enterprise-grade security.

Inverted Index Optimization

• Multi-match Query: Supports multi-field queries (multi-match), including configurations for fields, crossfields, minimum_should_match, etc.
• Similarity Algorithm Parameter Adjustment: Supports adjusting the k1 and b parameters for the BM25 similarity algorithm.

Ecosystem Integration

• New External Catalog Support: Lakehouse now supports EXTERNAL CATALOG compatible with the Iceberg REST protocol, enabling direct connection to services such as Snowflake Open Catalog.

Bug Fixes

SQL Engine Fixes

• `` Function: Fixed an issue regarding the return value of select if null.
• Dynamic Table Field Types: Added support for manually specifying field types when creating a Dynamic Table.
• Table Comment Replication: Fixed an issue where create table like did not replicate the comments from the source table.

Performance Optimization

• Metadata Performance: Column stats are no longer collected during imports in PIPE (COPY SQL) scenarios to improve COMMIT efficiency.
• Meta Cache: Added adaptive caching strategies for Meta Cache.
• HDFS Writing: Resolved Java client compatibility issues when writing to HDFS external tables.

User Experience Improvements

• information_schema: Fixed an issue where duplicate entries for tables with the same name were displayed in the information_schema.tables view.
• Job Profile: Enhanced job profile information display, now supporting the exposure of sub-job information.