IceFrame is a Python library that wraps Apache Iceberg with a clean, high-level API. It removes boilerplate from common tasks such as creating tables, reading and writing data, evolving schemas, managing partitions, branching, and rolling back snapshots. It also adds features the core Iceberg libraries do not provide, including compaction tools, partition evolution helpers, async operations, and a natural-language agent that can inspect schemas and generate Python code.
This talk shows how IceFrame streamlines daily work for engineers who build or maintain Iceberg pipelines and gives direct examples taken from real scripts and workflows.