Community Weekly Report
Oct 5 – Oct 11, 2026
00Overview
This week, Apache Doris is moving toward more complete lakehouse reads and writes, greater capacity for large queries, and stronger cloud ingestion and identity authentication, while bringing semi-structured data into relational queries. Kernel and engineering remain the center of community activity, followed by real-time data warehousing, with lakehouse and cloud-native capabilities also advancing. With 39 contributors participating, 76 PRs merged, and 38 new issues opened, the community shows sustained development and feedback activity.
01Highlights
02By Scenario
Multi-modal Lakehouse
01Lakehouse data needs consistent semantics across connectors, JDBC, file readers, and external writes. This week strengthened binary, UUID, and timestamp mappings. Affected binary fields now expose VARBINARY and preserve invalid UTF-8, embedded NULs, and trailing zeros, including nested fields in file TVFs, reducing conversion-related data distortion.
Lakehouse work uses metadata pruning, partition-value reads and scheduling improvements to avoid unnecessary file access and hotspots. Newly opened priorities cover AI connection reuse and optional filename padding, alongside external-write compatibility, aiming to reduce query and AI overhead and make exported filenames easier to sort.
- [improvement](hive) Push partition filters to HMS
- [improvement](hive) Support partition-column-value-only pushdown for Hive/Hudi scan
- [improvement](fe) Optimize external file COUNT scan scheduling
- [feature](outfile) add opt-in file_name_padding
- [improvement](be) Reuse HTTP connections across AI function requests
