Managing Snowflake Notebooks with Python¶

Prerequisites¶

The examples in this topic assume that you’ve added code to connect with Snowflake and to create a Root object from which to use the Snowflake Python APIs.

For example, the following code uses connection parameters defined in a configuration file to create a connection to Snowflake:

from snowflake.core import Root
from snowflake.snowpark import Session

session = Session.builder.config("connection_name", "myconnection").create()
root = Root(session)

Using the resulting Session object, the code creates a Root object to use the API’s types and methods. For more information, see Connect to Snowflake with the Snowflake Python APIs.

Creating a notebook¶

To create a notebook, first create a Notebook object, and then create a NotebookCollection object from the API Root object. Using NotebookCollection.create, add the new notebook to Snowflake.

Code in the following example creates a Notebook object that represents a notebook named my_nb in the my_db database and the my_schema schema:

from snowflake.core.notebook import Notebook

my_nb = Notebook(name="my_nb")

notebooks = root.databases["my_db"].schemas["my_schema"].notebooks
notebooks.create(my_nb)

The code creates a NotebookCollection variable notebooks and uses NotebookCollection.create to create a new notebook in Snowflake.

You can also create a notebook from a stage with an existing .ipynb file. Code in the following example creates a notebook from the @my_stage stage with the notebook_file.ipynb file:

from snowflake.core.notebook import Notebook

my_nb = Notebook(name="my_nb",
  query_warehouse="my_wh",
  from_location="@my_stage",
  main_file="notebook_file.ipynb")

notebooks = root.databases["my_db"].schemas["my_schema"].notebooks
notebooks.create(my_nb)

Getting notebook details¶

You can get information about a notebook by calling the NotebookResource.fetch method, which returns a Notebook object.

Code in the following example gets information about a notebook named my_nb in the my_db database and the my_schema schema:

my_nb = root.databases["my_db"].schemas["my_schema"].notebooks["my_nb"].fetch()
print(my_nb.to_dict())

Listing notebooks¶

You can list notebooks using the NotebookCollection.iter method, which returns a PagedIter iterator of Notebook objects.

Code in the following example lists notebooks whose name starts with my in the my_db database and the my_schema schema, and then prints the name of each:

from snowflake.core.notebook import NotebookCollection

notebooks: NotebookCollection = root.databases["my_db"].schemas["my_schema"].notebooks
nb_iter = notebooks.iter(like="my%")  # returns a PagedIter[Notebook]
for nb_obj in nb_iter:
  print(nb_obj.name)

Performing notebook operations¶

You can perform common notebook operations—such as managing versions and executing notebooks—with a NotebookResource object.

To demonstrate some operations you can do with a notebook resource, code in the following example does the following:

1. Gets the my_nb notebook resource object.

2. Adds a lives version to the notebook object. This is equivalent to ALTER NOTEBOOK … ADD LIVE VERSION.

3. Commits the live version of the notebook to a Git repository, if a Git connection is set up. Otherwise, sets the live version to null.

   For more information, see ALTER NOTEBOOK.

4. Executes the notebook.

   Note

   To execute a notebook, you must add a live version to it first.

5. Drops the notebook.

my_nb_res = root.databases["my_db"].schemas["my_schema"].notebooks["my_nb"]

my_nb_res.add_live_version(from_last=True)
my_nb_res.commit()
my_nb_res.execute()
my_nb_res.drop()