Query Apache Iceberg Tables

• Unity native Iceberg (REST): Not supported.

  Workaround: Use Delta + UniForm to publish an Iceberg-readable view via Unity Catalog’s Iceberg REST endpoint.

• Certified REST catalogs: ADB is certified with Snowflake Polaris and Databricks Unity Catalog (UniForm only) for Iceberg read access.

• Catalog credential vending: Not supported.

  Cloud-native role-based vending, such as automatic role assumption or temporary credentials issued by STS, is not supported. Use explicit access/secret keys or static tokens.)

• AWS ARN credentials: Not supported. IAM role ARNs and AssumeRole via ARN are not accepted.

• Partitioned Iceberg tables are not supported; only non-partitioned tables are allowed.
• Row-level updates (merge-on-read): Not supported. If Iceberg metadata references delete files, queries will fail.

• The schema of a fixed external table is determined at creation and must align with the Iceberg schema version in the metadata. If the Iceberg schema is updated, the external table must be recreated.

• No query-time time travel: Querying by snapshot, version, or timestamp is not supported.
• Non_catalog only: New snapshots are not picked up automatically. To read a specific snapshot, target that snapshot’s metadata.json and recreate the external table.

• Credential alignment: The same credentials must be used for AWS S3 and AWS Glue.
• Region co-location: S3 buckets and the AWS Glue catalog must be in the same AWS region.

• A Delta table created with UniForm so Iceberg clients can read it.
• Table files in Azure ADLS Gen2 or AWS S3.
• Unity Catalog privileges for external access (e.g., External data access enabled); grant EXTERNAL USE SCHEMA to your principal).
• Auth: OAuth2 (recommended) or Personal Access Token (for quick tests).

USE CATALOG <your_catalog>;
USE SCHEMA  <your_schema>;

CREATE TABLE customers_iceberg (
  id   INT,
  name STRING
)
TBLPROPERTIES(
  'delta.columnMapping.mode'='name',
  'delta.enableIcebergCompatV2'='true',
  'delta.universalFormat.enabledFormats'='iceberg'
);

INSERT INTO customers_iceberg (id, name) VALUES
  (1,'Alice'), (2,'Bob'), (3,'Carol');

BEGIN
  BEGIN DBMS_CLOUD.DROP_CREDENTIAL('AZURE_BLOB_CRED'); EXCEPTION WHEN OTHERS THEN NULL; END;
  DBMS_CLOUD.CREATE_CREDENTIAL(
    credential_name => 'AZURE_BLOB_CRED',
    username        => '<storage-account-or-sas-username>',
    password        => '<storage-key-or-sas-token>'
  );
END;
/

-- Databricks service principal (client_id / client_secret)
BEGIN
  BEGIN DBMS_CLOUD.DROP_CREDENTIAL('UNITY_OAUTH'); EXCEPTION WHEN OTHERS THEN NULL; END;
  DBMS_CLOUD.CREATE_CREDENTIAL(
    credential_name => 'UNITY_OAUTH',
    username        => '<client_id>',
    password        => '<client_secret>'
  );
END;
/

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE(
    table_name      => 'CUSTOMERS_ICEBERG',
    credential_name => 'AZURE_BLOB_CRED',
    format          => '{
      "access_protocol": {
        "protocol_type": "iceberg-rest",
        "protocol_config": {
          "iceberg_catalog_type": "unity",
          "rest_catalog_endpoint": "https://<workspace-host>/api/2.1/unity-catalog/iceberg",
          "rest_authentication": {
            "rest_auth_cred": "UNITY_OAUTH",
            "rest_auth_endpoint": "https://<workspace-host>/oidc/v1/token"
          },
          "table_path": ["<your_catalog>","<your_schema>","customers_iceberg"]
        }
      }
    }'
  );
END;
/
SELECT COUNT(*) FROM CUSTOMERS_ICEBERG;

BEGIN
  BEGIN DBMS_CLOUD.DROP_CREDENTIAL('UNITY_PAT'); EXCEPTION WHEN OTHERS THEN NULL; END;
  DBMS_CLOUD.CREATE_CREDENTIAL(
    credential_name => 'UNITY_PAT',
    username        => 'token',
    password        => '<dapiXXXXXXXXXXXXXXXXXXXXXXXX>'
  );
END;
/

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE(
    table_name      => 'CUSTOMERS_ICEBERG',
    credential_name => 'AZURE_BLOB_CRED',
    format          => '{
      "access_protocol": {
        "protocol_type": "iceberg-rest",
        "protocol_config": {
          "iceberg_catalog_type": "unity",
          "rest_catalog_endpoint": "https://<workspace-host>/api/2.1/unity-catalog/iceberg",
          "rest_authentication": { "rest_auth_cred": "UNITY_PAT" },
          "table_path": ["<your_catalog>","<your_schema>","customers_iceberg"]
        }
      }
    }'
  );
END;
/
SELECT COUNT(*) FROM CUSTOMERS_ICEBERG;

• Polaris Iceberg catalog and endpoint available to your account.
• Table files accessible in your object store (S3/ADLS as applicable).
• Auth: OAuth2 recommended (client credentials) or another Polaris-supported token mechanism.

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE(
    table_name      => 'SALES_POLARIS',
    credential_name => 'S3_CRED',
    format          => '{
      "access_protocol": {
        "protocol_type": "iceberg-rest",
        "protocol_config": {
          "iceberg_catalog_type": "polaris",
          "rest_catalog_endpoint": "<https://<your-polaris-endpoint>/...>",
          "rest_authentication": {
            "rest_auth_cred": "POLARIS_OAUTH",
            "rest_auth_endpoint": "<https://<your-oauth-token-endpoint>>"
          },
          "table_path": ["<db>","<schema>","<table>"]
        }
      }
    }'
  );
END;
/

SELECT COUNT(*) FROM SALES_POLARIS;

• Glue Data Catalog with the Iceberg table registered (S3 objects accessible).

  Region name for Glue (e.g., us-east-1).

• Auth: Access key/secret for S3; Glue access via catalog configuration.

• Credential vending via AWS ARN is unsupported. Explicit credentials must be supplied.

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE(
    table_name      => 'ORDERS_GLUE',
    credential_name => 'S3_CRED',
    format          => '{
      "access_protocol": {
        "protocol_type": "iceberg",
        "protocol_config": {
          "iceberg_catalog_type": "aws_glue",
          "iceberg_glue_region": "us-east-1",
          "table_path": ["<database>","<table>"]
        }
      }
    }'
  );
END;
/

SELECT COUNT(*) FROM ORDERS_GLUE;

• You can access the table’s Iceberg root manifest (metadata.json) in your object store (ADLS/S3/OCI).
• This path is point-in-time. To follow new snapshots, recreate the external table.

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE(
    table_name      => 'CUSTOMERS_META',
    credential_name => 'STORE_CRED',
    file_uri_list   => 'https://<bucket-or-container>/<path>/metadata.json',
    format          => '{"access_protocol":{"protocol_type":"iceberg"}}'
  );
END;
/

BEGIN
  DBMS_CLOUD.CREATE_EXTERNAL_TABLE (
    table_name       => 'iceberg_parquet_time_dim3',
    credential_name  => 'OCI_CRED',
    format           => '{
      "access_protocol": {
        "protocol_type": "iceberg",
        "protocol_config": {
          "iceberg_catalog_type": "hadoop",
          "iceberg_warehouse": "https://objectstorage.uk-cardiff-1.oraclecloud.com/n/my-tenancy/b/my-iceberg-bucket/o/iceberg",
          "iceberg_table_path": "db.icebergTablePy"
        }
      }
    }'
  );
END;
/