The Next Era of the Open Lakehouse: Apache Iceberg™ v3 in Public Preview on Databricks

Databricks announces Apache Iceberg v3 entering Public Preview on its platform. The new version introduces three major features: Row Lineage (permanent row IDs and sequence numbers for efficient change tracking), Deletion Vectors (lightweight delete files that make data manipulation up to 10x faster than copy-on-write), and the VARIANT type (native semi-structured data storage queryable with standard SQL). These features eliminate the historical tradeoff between performance and cross-engine interoperability. Unity Catalog serves as the governance hub, supporting fine-grained access controls enforced across all engines. Delta Lake with UniForm allows writing once and reading as Iceberg from Snowflake, BigQuery, Redshift, Athena, and others. Iceberg v3 is available on Databricks Runtime 18.0+ with Unity Catalog enabled. Databricks engineers are also driving Iceberg v4 proposals including an adaptive metadata tree, relative path support, and a modernized statistics model.