Apache Iceberg
Connect Apache Iceberg tables through a catalog. Iceberg tables are mounted into the query engine and can be joined with datasets from other connectors.
How Iceberg is addressed
An Iceberg connection has three distinct parts. Keeping them separate is the key to configuring it correctly:
- Catalog — a metadata service that tracks tables, schemas, and snapshots. Datyo talks to a REST catalog (Tabular, Polaris, Lakekeeper, Nessie).
- Warehouse — an identifier registered in the catalog, not a filesystem path. The catalog resolves it to the actual storage location. (A placeholder such as
prod-warehouseis a typical value.) - Storage — the object storage where the table's data files physically live. The catalog returns these locations; Datyo reads the files using the storage credentials you provide.
The Catalog URI is the metadata service address; the Warehouse names a warehouse inside that catalog; the storage credentials are the keys used to read the data files it points to. These are three different things, not three ways of writing the same location.
Before you begin
- The address and token of your REST catalog service.
- Read credentials for the object storage that backs the warehouse (see object-storage access credentials).
Connection settings
| Field | Required | Description |
|---|---|---|
| Catalog type | Yes | REST (Tabular / Polaris / Lakekeeper / Nessie). |
| Catalog URI | Yes | The REST catalog service address, e.g. https://catalog.example.com. |
| Warehouse | Yes | The warehouse identifier registered in the catalog — not a path, e.g. prod-warehouse. |
| Token | — | Bearer token for the catalog, if it requires authentication. Stored encrypted. |
Then provide the object-storage access credentials for the storage that holds the data files.
Test and save
Enter the fields and use Test connection. Datyo contacts the REST catalog, then reads table metadata and a sample of data via the storage credentials. On success, choose the tables to import.
Capabilities
- Execution model: Federated. Iceberg tables can be joined with datasets from other connectors.
- Access: Read-only.
- Catalog support: REST catalogs are supported today. AWS Glue and S3 Tables catalogs are shown in the console as planned and are not yet available.
Troubleshooting
- Cannot reach the catalog — verify the Catalog URI and that the catalog service is reachable from Datyo; provide a Token if the catalog requires authentication.
- Tables list but data cannot be read — the catalog resolved, but the storage credentials are missing or lack read access to the data files; check the object-storage keys and region.
- "Warehouse not found" — the Warehouse value must be the identifier registered in the catalog, not a bucket path.