Databend's January monthly report covers 11 nightly releases (v1.2.863 to v1.2.873-nightly) with 33 new features, 31 bug fixes, and 37 refactors. Highlights include experimental Table Branching & Tags for data version control, full Spatial Index support for Geometry/Geography workloads, and newly live Iceberg Write support (via iceberg-rust v0.8.0). Engine improvements include self-join elimination, lazy materialization across joins, and Fuse Parquet Prewhere optimizations, alongside Meta service observability and Openraft upgrades. A spotlight section frames Table Branching as a safe sandbox for AI agents to experiment on data without touching production tables.
Table of contents
By the NumbersMonthly HighlightsSpotlight: Table Branching & Tag (The Perfect Data Sandbox for Agents)Questions this post answers
What is Table Branching and Tags in Databend used for?
Table Branching and Tags is an experimental Databend feature enabling data version control by letting users create branches of a table for safe experimentation. Agents or users can perform DELETE, UPDATE, or complex ETL operations on a branch without affecting the main table, then merge verified changes back or use tags for snapshot-based rollback via time travel. Teams evaluating data version control workflows can follow Databend's evolving feature set on daily.dev.
Does Databend support writing to Iceberg tables?
Yes, Databend added Iceberg write support by upgrading to iceberg-rust v0.8.0, enabling the write path in addition to existing read support and improving compatibility with open data lake formats. This was delivered as part of a run of 11 nightly releases spanning versions v1.2.863 through v1.2.873. Developers weighing lakehouse write compatibility across engines can track updates like this on daily.dev.
What spatial indexing capabilities does Databend offer for Geometry and Geography data?
Databend introduced Spatial Index support with CREATE/DROP SPATIAL INDEX statements that fully cover Geometry and Geography workloads, boosting spatial query performance. This was shipped alongside other query engine improvements like self-join elimination and lazy materialization across joins during the same nightly release cycle. Engineers building geospatial analytics pipelines can follow database feature rollouts like this on daily.dev.