A 2026 Introduction to Apache Iceberg

Apache Iceberg is an open-source table format for large analytic datasets. It defines how data files stored on object storage (S3, ADLS, GCS) are organized into a logical table with a schema, partition layout, and consistent point-in-time snapshots. If you've heard the term "data lakehouse," Iceberg is the layer that makes it possible by bringing warehouse-grade reliability to data lake storage. This post covers what Iceberg is, how its metadata works under the hood, what changed across specification versions 1 through 3, what's being proposed for v4, and how to get started using Iceberg tables with Dremio in about ten minutes. Where Iceberg Came From Before Iceberg, most data lake tables used the Hive table format. Hive tracks data by directory paths: one directory per partition, with files inside. That works fine for small tables, but it breaks down at scale. Listing files across thousands of partition directories takes minutes. Schema changes require careful coordination. There's no