May 19-20, 2025 — 9.13 Release Notes¶

Outbound private connectivity for AWS Government regions¶

You can use AWS PrivateLink for outbound network traffic originating from AWS Government regions.

For information about using AWS PrivateLink for outbound network traffic, see Manage private connectivity endpoints: AWS.

Pipe operator¶

With this release, you can use the new pipe operator (->>) to chain SQL statements together. In the chain of SQL statements, the results of one statement can serve as the input to another statement. The pipe operator can simplify the execution of dependent SQL statements and improve the readability and flexibility of complex SQL operations.

For more information, see Flow operators.

INFER_SCHEMA function: Support for Apache Iceberg™ data types¶

With this release, you can now automatically retrieve the column definitions for Apache Iceberg data types from a set of staged data files that contain semi-structured data. To retrieve these column definitions, set the new KIND argument for the INFER_SCHEMA function to ICEBERG.

With this new feature, you can use the CREATE ICEBERG TABLE … USING TEMPLATE variant syntax to create an Iceberg table with the column definitions derived from INFER_SCHEMA using the KIND=>'ICEBERG' output.

To learn more, see INFER_SCHEMA.

Cross-cloud/cross-region support for Snowflake-managed Apache Iceberg™ tables¶

With this release, Snowflake supports cross-cloud/cross-region writes (and reads) to Iceberg tables that use Snowflake as the catalog. In the DATA_TRANSFER_HISTORY view, the cross-region data transfer charges for these tables appear as a TRANSFER_TYPE of DATA_LAKE. For more information, see the following:

• Cross-cloud/cross-region support

• DATA_TRANSFER_HISTORY view in the ORGANIZATION_USAGE schema

• DATA_TRANSFER_HISTORY view in the ACCOUNT_USAGE schema

In addition, you can now convert a cross-cloud/cross-region Iceberg table that uses an external catalog to use Snowflake as the catalog. For more information, see Convert an Apache Iceberg™ table to use Snowflake as the catalog.