4. Working with Iceberg Tables

This chapter describes one Apache Iceberg table model and its Oracle access paths.Oracle can read an existing Iceberg table from explicit metadata or through a supported mounted catalog. Where the catalog, storage, deployment, and release support it, Oracle can also create and publish data to the same catalog-managed Iceberg table model.

A mounted catalog is the common operational boundary: it resolves current table metadata for existing tables and enables Oracle-managed create/write operations. Native Iceberg in the below table identifies Oracle’s managed create/write path; it is not a different table format or a table type.

Flow: One Iceberg Table, Two Oracle Workflows

Mount the authoritative Catalog	→	Read an existing Table	→	Optional Oracle-managed Publish	→	Commit becomes visible
DBMS_CATALOG resolves names and current metadata		Discover → resolve snapshot → plan manifests → read selected files		Create or insert through the same supported catalog path		Catalog pointer advances; external engines read the shared table state

Complete the common package, network, identity, and credential setup as explained in Chapter 2. This chapter does not repeat those prerequisites.