Snowflake Notebooks: Notebook Project Objects are now Code Bundles (Pending)¶
Notebook Project Objects (NPOs) have been renamed to Code Bundles. The object, its behavior, and its access model are unchanged: only the name and the corresponding SQL, CLI, and Snowsight labels are updated. Existing objects are recognized as Code Bundles automatically, and no migration is required.
This rename is the foundation for a broader capability. The name Code Bundle reflects a more general object for packaging and running non-SQL code on Snowflake compute, and it sets up expanded functionality and additional runtime options planned for future releases. Renaming the existing, generally available capability now lets your workflows adopt the new name early, with no disruption, ahead of those additions.
Your existing Notebook Projects usage keeps working. EXECUTE NOTEBOOK PROJECT statements, including those inside existing tasks and schedules, continue to run without changes, and CREATE NOTEBOOK PROJECT still creates the object. Code Bundles are the same object under a new name, not a replacement you have to migrate to.
- Before the change:
You created and ran these objects with the
NOTEBOOK PROJECTgrammar, and Snowsight listed them as Notebook Projects:- After the change:
The same objects are Code Bundles, created and run with the
CODE BUNDLEgrammar, and Snowsight lists them as Code Bundles:The
NOTEBOOK PROJECTgrammar shown before the change keeps working.EXECUTE NOTEBOOK PROJECTcontinues to be supported, so existing schedules never break.
Actions required¶
None. This change is backward compatible:
- Existing objects are recognized as Code Bundles automatically.
- Existing
EXECUTE NOTEBOOK PROJECTandCREATE NOTEBOOK PROJECTstatements continue to work. - Existing tasks and schedules continue to run without changes.
We recommend using the CODE BUNDLE grammar for new objects.
What Code Bundles unlock¶
The rename is the foundation for an expanded capability. Beyond the notebook-scheduling and compute pool (Snowpark Container Services) execution that is generally available (the capability previously delivered as Notebook Projects), Code Bundles introduce the following in Public Preview:
- Warehouse execution: run Python Code Bundles directly on warehouse compute, in addition to compute pools.
- Spark jobs: submit Spark applications (Scala, Java, or Python) as batch jobs on Snowflake compute. See Submit Spark jobs on Snowflake.
- Inline specification overrides: pass configuration at execution time with
WITH SPECIFICATION, instead of relying only on the storedcode_bundle.yml. - New clients: the REST API, Python API, and Snowflake CLI for creating and executing Code Bundles.
Together, these establish Code Bundles as a general foundation for packaging and running non-SQL code on Snowflake compute, without building containers or wrapping logic in stored procedures. For the full guide, see Snowflake Code Bundles.
Additional notes¶
- For details about creating, running, and scheduling Code Bundles, see Run and schedule Notebooks in Workspaces.
- For the SQL reference, see CREATE CODE BUNDLE, EXECUTE CODE BUNDLE, and SHOW CODE BUNDLES.
- The
default_version_location_urivalue returned byDESCRIBE CODE BUNDLEandDESCRIBE NOTEBOOK PROJECTnow uses thesnow://code bundle/...scheme (previouslysnow://notebook project/...).
Ref: 2393