Preview — 9.8 Release Notes¶

Trust Center: Risky Human and Service User scanners¶

With this release, Snowflake is pleased to announce two additional scanners: Human User MFA Readiness and Service User Passwordless Readiness. As part of the Threat Intelligence scanner package, the new scanners allow you to check for risky human and service users to further reduce security vulnerabilities.

• Human User MFA Readiness Scanners identify human users who have signed in with just a password in the last 90 days and haven’t yet set up multi-factor authentication (MFA). It also flags human users who haven’t logged in in 90 days but still have a password set.

• Service User Passwordless Readiness looks for Legacy service users who have recently logged in with a password and haven’t removed it. It also flags service users who haven’t logged in in 90 days but still have a password set.

Bind variables in SHOW commands¶

With this release, you can use bind variables with the LIKE and LIMIT keywords in SHOW commands. For example, the following SHOW command, which could be included in a Javascript stored procedure, uses bind variables:

SHOW TABLES LIKE ? LIMIT ?;

The following example uses bind variables in a SHOW command in a Snowflake Scripting block:

BEGIN
  LET a INT := 10;
  LET p STRING := 'mytable';
  LET res RESULTSET := (SHOW TABLES LIKE :p LIMIT :a);
  RETURN TABLE(res);
END;

Apache Iceberg™ tables: Row-level deletes for externally managed tables — Preview¶

With this release, we are pleased to announce the preview of row-level deletes support with positional delete files when external engines perform update, delete, and merge operations on externally managed Iceberg tables in Snowflake.

Apache Iceberg™ tables: Delta table support — General availability¶

With this release, we are pleased to announce general availability support for creating read-only Iceberg tables from Delta Lake tables stored in object storage. Creating Iceberg tables sourced from Delta Lake delta logs provides the ability to perform efficient Lakehouse analytics in Snowflake and generate Iceberg metadata for consumption in an Iceberg engine ecosystem.

New database properties: CATALOG_SYNC_NAMESPACE_MODE and CATALOG_SYNC_NAMESPACE_FLATTEN_DELIMITER¶

With this release, Snowflake is pleased to announce the release of two new database properties:

• CATALOG_SYNC_NAMESPACE_MODE

• CATALOG_SYNC_NAMESPACE_FLATTEN_DELIMITER

This allows you to query Snowflake-managed Apache Iceberg™ tables in Open Catalog by using a third-party engine that can only query tables grouped up to two levels deep in the namespace hierarchy, such as Trino.

Use the flatten setting for the CATALOG_SYNC_NAMESPACE_MODE property to sync a Snowflake-managed Iceberg table to Snowflake Open Catalog with one parent namespace by flattening the table’s two parent namespaces into one namespace. Use the CATALOG_SYNC_NAMESPACE_FLATTEN_DELIMITER property to insert a delimiter in the resulting namespace to avoid conflicts that could arise from flattening the two parent namespaces. You specify these properties when you create a database.

For example, when you set the CATALOG_SYNC_NAMESPACE_MODE property to flatten and specify an en-dash (-) for the CATALOG_SYNC_NAMESPACE_FLATTEN_DELIMITER property, Snowflake syncs the customer.data.table1 and custom.erdata.table1 Snowflake-managed Iceberg tables to the catalog1 external catalog in Open Catalog with the following fully qualified names:

• catalog1.customer-data.table1

• catalog1.custom-erdata.table1

If you use the default for the CATALOG_SYNC_NAMESPACE_MODE property (nest), Snowflake continues to sync the table to Open Catalog with two parent namespaces and the CATALOG_SYNC_NAMESPACE_FLATTEN_DELIMITER property isn’t required.