Apache Iceberg

Apache Iceberg is an open table format for analytic datasets on object storage. It layers table semantics — ACID transactions, snapshot isolation, schema and partition evolution, time travel — over files in formats such as Parquet.

Iceberg is engine-agnostic by design: Spark, Trino, Flink, and a growing set of warehouses and managed platforms can read and write the same Iceberg tables through a catalog. Hidden partitioning removes the need for query authors to know the physical layout.

It is a foundational component of lakehouse architectures, chosen when multiple engines must share one copy of the data. The main comparable format is Delta Lake, with interoperability between formats an active area of the ecosystem.