Apache Iceberg C++ 0.4.0 Release

The Apache Iceberg community is pleased to announce the 0.4.0 release of Apache Iceberg C++. This release includes over 160 pull requests from 26 contributors, including 12 first-time contributors.

iceberg-cpp is a native C++ implementation of the Apache Iceberg table format, providing libraries for reading, writing, and managing Iceberg tables in C++ applications.

Release Highlights

Iceberg v3 Support

• v3 type definitions and v3 geometry and geography support
• Column default values end to end: representation, serialization, and validation, Parquet reads of missing fields, Avro reads of missing fields, and UpdateSchema support
• Row lineage: writing snapshot row lineage fields at the top level, reading row lineage metadata columns, and retrying stale row-lineage validation
• Deletion vectors: deletion-vector-v1 blob read and write, merging multiple DVs in MergingSnapshotUpdate, and Puffin and DV support moved into the core library

Table Update APIs

• New snapshot-producing operations: DeleteFiles, RowDelta, OverwriteFiles, RewriteFiles, ReplacePartitions (now exposed on Table and Transaction), and merge append
• Retryable and cleanup-safe table updates, FileCleanupStrategy hardened with retries and parallel deletes, and cleanup of delete files from expired manifests
• Catalog operations including RenameTable in InMemoryCatalog and purge support in DropTable for the in-memory and SQL catalogs

Catalogs and FileIO

• REST catalog gains SigV4 authentication, a session-aware catalog, OAuth2 token exchange sessions, and per-table FileIO bound from vended storage credentials
• ResolvingFileIO selects a FileIO by location scheme and forwards vended credentials, backed by a registry-driven resolution, with S3-compatible schemes and the oss scheme supported
• Hive Metastore groundwork with vendored HMS IDL and generated bindings, an iceberg_hive library, and HmsClient connection lifecycle and URI parsing
• Closer alignment with the Java implementation for REST table update serialization and REST error handling

Scan Planning and Performance

• Lazy scan planning streams and a fallible iterator utility
• Parallelism for reading manifests, writing manifests, and update and scan processing, plus an LRU cache
• Metadata tables with a base metadata table interface and streaming SnapshotsTable scans
• Arrow data access via ArrowRowBuilder and reading list columns as large_list
• Metrics with commit and scan reporting integration, Parquet NaN metrics collected during writes, and a unique-value optimization for notEq/notIn

Logging

• A pluggable logging stack built up across the release: LogLevel, the Logger interface and default logger, a std::cerr backend, logging macros, an optional spdlog backend behind ICEBERG_SPDLOG, and a Loggers registry
• Transaction commit lifecycle logging as the first consumer

Build Changes

• Meson build support has been removed; CMake is now the only supported build system
• Dependency upgrades to Arrow 25.0.0 and nanoarrow 0.9.0

Contributors

$ git shortlog --perl-regexp --author='^((?!dependabot\[bot\]).*)$' -sn v0.3.0..v0.4.0
    23  Junwang Zhao
    14  Gang Wu
    13  Manu Zhang
    13  Zehua Zou
    13  wzhuo
    10  Abanoub Doss
    10  Minh Vu
     9  Xin Huang
     8  kamcheungting-db
     7  Jiajia Li
     5  YangJie
     3  Kevin Liu
     3  Rahul Goel
     3  Xinli Shang
     3  lishuxu
     3  liuxiaoyu
     2  Anupam Yadav
     2  Joey
     2  Yuya Ebihara
     1  Huangshi Tian
     1  Innocent Djiofack
     1  Rahul Shivu Mahadev
     1  Sreesh Maheshwar
     1  Timothy Wang
     1  ZhaoXuan
     1  kid

This release welcomes 12 first-time contributors to Apache Iceberg C++: @kamcheungting-db, @All-less, @goel-skd, @abnobdoss, @ebyhr, @huan233usc, @yadavay-amzn, @timothyw553, @LuciferYang, @Angelia-Wang, @rahulsmahadev, and @u70b3.

We thank all contributors for their efforts in making this release possible!

Roadmap for 0.5.0

The community is tracking the next release in #959, which focuses on the remaining Iceberg v3 feature support.

Getting Involved

We welcome questions and contributions from all interested. Issues can be filed on GitHub, and questions can be directed to GitHub or the Iceberg dev mailing list.