Configure a catalog integration for files in object storage¶

Create a catalog integration for Apache Iceberg™ table files, Delta Lake table files, or Apache Parquet™ files in object storage.

After you create a catalog integration, you can create an Iceberg table.

Iceberg files¶

Create a catalog integration for Iceberg metadata that’s in an external cloud storage location by setting OBJECT_STORE as the CATALOG_SOURCE value and ICEBERG as the TABLE_FORMAT.

CREATE OR REPLACE CATALOG INTEGRATION icebergCatalogInt
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = ICEBERG
  ENABLED = TRUE;

Delta Lake table files¶

This integration is part of Delta Direct in Snowflake: you use it together with CREATE ICEBERG TABLE (Delta files in object storage) so that Snowflake can read Delta Lake tables in your bucket and optionally generate Iceberg metadata.

Create a catalog integration for Iceberg tables based on Delta table files by setting OBJECT_STORE as the CATALOG_SOURCE value and DELTA as the TABLE_FORMAT.

• CATALOG_SOURCE = OBJECT_STORE
• TABLE_FORMAT = DELTA

CREATE OR REPLACE CATALOG INTEGRATION delta_catalog_integration
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = DELTA
  ENABLED = TRUE;

Parquet files¶

This integration is part of Parquet Direct in Snowflake: you use it together with CREATE ICEBERG TABLE (Parquet files in object storage) so that Snowflake can read Parquet files in your bucket in place, without a separate table format. For more information, see Parquet Direct.

Set OBJECT_STORE as the CATALOG_SOURCE value, and choose the TABLE_FORMAT that matches how your files are organized:

• TABLE_FORMAT = HIVE: Your files use Hive-style key=value directory partitioning (for example, year=2024/month=01/data.parquet). Snowflake infers partition columns from the paths.
• TABLE_FORMAT = NONE: Your files use any other layout. You can define partition columns explicitly by using expressions on the METADATA$FILENAME pseudocolumn.

-- Hive-style partitioned Parquet files
CREATE OR REPLACE CATALOG INTEGRATION hive_catalog_int
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = HIVE
  AUTO_REFRESH = TRUE
  ENABLE_SCHEMA_EVOLUTION = TRUE;

-- Parquet files with a custom directory layout
CREATE OR REPLACE CATALOG INTEGRATION parquet_catalog_int
  CATALOG_SOURCE = OBJECT_STORE
  TABLE_FORMAT = NONE
  ENABLED = TRUE
  AUTO_REFRESH = TRUE
  ENABLE_SCHEMA_EVOLUTION = TRUE;