December 03-05, 2024 — 8.45 Release Notes¶

Snowflake Scripting: Asynchronous child jobs — Preview¶

With this release, Snowflake Scripting (SQL) stored procedures can run queries concurrently as asynchronous child jobs. The query can be any valid SQL statement, including SELECT statements and DML statements, such as INSERT or UPDATE.

To run a query as an asynchronous child job, add the ASYNC keyword to the query for a RESULTSET. When this keyword is omitted, the stored procedure runs child jobs sequentially, and each child job waits for the running child job to finish before it starts.

Running multiple child jobs concurrently can improve efficiency and reduce overall run time.

For more information, see Assigning a query to a declared RESULTSET.

Profiling Python stored procedure handlers — Preview¶

With this release, you can discover how much time or memory was spent executing your stored procedure handler code by using the built-in code profiler. The profiler generates information describing how much time or memory was spent executing each line of Python code.

For more information, see Profiling Python procedure handler code (for SQL API) and Profiling Snowpark Python stored procedure handlers (for Python API).

Java 17 support — General Availability¶

With this release, we are pleased to announce general availability of Java 17 in Snowpark. You can now create and run stored procedures and UDFs using Java 17. The Snowpark API and JDBC Driver have also been updated to support Java 17.

For more information, see Snowflake Java Runtime Support.

Dynamic tables: Unlimited inputs¶

With this release, you can define dynamic tables that read from an unlimited number of tables or dynamic tables. Previously, dynamic table definitions were limited to querying up to 100 tables or dynamic tables.