PhoenixAI 26.2: More Workloads on Shared-Data and Iceberg

🎟 Join us for NEXT QUERY 2026 | Free | Virtual | Nov 4 👉 20+ practical talks from teams building and operating data + AI infrastructure in production: https://bit.ly/4hBgFGR PhoenixAI 26.2 lets you run more workloads on the data you already have: vector search and incremental materialized views on shared-data clusters, UPDATE and MERGE INTO on Apache Iceberg tables, and much less manual tuning. In this walkthrough, Sida Shen covers what's new, why it matters, and what to check before you upgrade from StarRocks 4.1. If you're serving multi-tenant analytics, keeping materialized views fresh on large tables, or still running Spark jobs just to update Iceberg tables, this release removes a lot of that extra work. What you'll learn: - How range distribution and automatic tablet split/merge handle data skew in multi-tenant tables - Vector search on shared-data clusters, including filtered vector search that applies predicates during the search - Incremental materialized view refresh: processing only the rows that changed instead of recomputing whole partitions - Iceberg UPDATE and MERGE INTO, v3 deletion vectors, VARIANT shredding, and automatic metadata maintenance - Upgrading from StarRocks 4.1 to 26.2, rolling back safely, and which defaults change PhoenixAI builds on StarRocks and stays compatible with the open-source version, adding enterprise security, high availability, and 24/7 support. 26.2 is also the first PhoenixAI release with a year-based version number: 26 stands for 2026. ---------------------------------------------------------------------------------------- Timestamps 0:00 Intro and Agenda 1:40 PhoenixAI and StarRocks: how they relate 2:49 Architecture: multi-warehouse, primary key tables, Apache Iceberg 5:00 Deployment: PhoenixAI Cloud (BYOC) and PhoenixAI Anywhere on Kubernetes 6:25 Enterprise features: column masking, row access policies, audit 7:27 PhoenixAI 26.2 version numbering and release cadence 9:30 Multi-tenant tables and data skew 11:02 Range distribution with automatic tablet split and merge 13:11 Range distribution improvements in 26.2 13:54 Vector search on shared-data clusters 15:13 Filtered vector search: before, after, and during search 15:54 Global late materialization in the cost-based optimizer 16:48 Synchronous vs. asynchronous materialized views 18:00 Incremental materialized view refresh 22:52 Apache Iceberg row-level updates 23:51 UPDATE and MERGE INTO on Iceberg tables (moving off Spark) 24:09 Iceberg v3 deletion vectors 25:03 Iceberg v3 VARIANT shredding 25:56 Automatic Iceberg metadata maintenance 27:02 Upgrading from StarRocks 4.1 to 26.2 (and rolling back) 27:54 JDBC aggregation pushdown, 2 GB VARCHAR, FILES() schema 29:33 Low-cardinality dictionary, cardinality estimation, faster primary key deletes 30:50 What's next: vector search and incremental MV deep dives ---------------------------------------------------------------------------------------- Learn more at https://www.phoenixdata.ai/ Connect with us: LinkedIn: https://www.linkedin.com/company/phoenixai-data/ Twitter: https://x.com/phoenixdataai CelerData Website: https://www.phoenixdata.ai/ StarRocks GitHub: https://github.com/StarRocks/StarRocks StarRocks Website: https://www.starrocks.io/ Slack: https://starrocks.io/redirecting-to-slack #RealTimeAnalytics #RealTimeData #DataEngineering #DataInfrastructure #DataArchitecture #OLAP #Database #ApacheIceberg #Lakehouse #StarRocks #OpenSource #StreamingData #ApacheKafka #CDC #DataEngineer