Ingest data in any format without upfront schema definition. Apply structure at query time for maximum flexibility and rapid iteration.

Native Apache Spark connector for distributed processing. Run PySpark, SparkSQL, and Spark ML directly against lake data at scale.

First-class support for Apache Parquet and ORC columnar formats. Predicate pushdown and column pruning for efficient analytical queries.

Integrated metadata catalogue with automatic schema detection. Tag, search, and discover datasets across departments and projects.