What problem does it solve?
Apache Iceberg on Databricks presents multiple deployment patterns (Managed Iceberg, UniForm external reads, Compatibility Mode, and IRC-based external access). This Skill describes how to deploy and interoperate Iceberg across Unity Catalog and external engines, including Snowflake facilitation and PyIceberg/OSS Spark integration.
Core Features & Use Cases
- Native Managed Iceberg tables with full read/write in Databricks.
- UniForm: making Delta tables readable as Iceberg by external engines without migrating data.
- Compatibility Mode for streaming tables and materialized views.
- Iceberg REST Catalog (IRC) enabling external engines to access UC-managed Iceberg data.
- Iceberg v3: advanced features such as deletion vectors and VARIANT types (beta).
- Snowflake interop: federated catalog access and cross-platform reads.
- PyIceberg and OSS Spark clients for external tooling and analytics.
- Credential vending and EXTERNAL USE SCHEMA grants for secure external access.
Quick Start
Create a managed Iceberg table in Unity Catalog or enable UniForm on a Delta table, then expose it to external engines via IRC and federated Snowflake integrations.