Choose a notebook image
Pick the JupyterLab base image for a dev environment, including which images run Spark.
When you launch a JupyterLab dev environment, you select a base image from the Custom Image dropdown. The image determines the Python version, the preinstalled packages, and whether you can run Spark from the notebook.
Available JupyterLab images
| Image name ends with | Package environment | Spark |
|---|---|---|
jupyterlab:python3.12 (default) | Python 3.12 | Yes |
jupyterlab:17.3-lts | Databricks Runtime 17.3 LTS | Yes |
jupyterlab:python3.13 | Python 3.13 | No |
jupyterlab:python3.14 | Python 3.14 | No |
jupyterlab:16.4-lts | Databricks Runtime 16.4 LTS | No |
Which image to use
- Start with the default.
jupyterlab:python3.12runs Spark DataFrames, Spark SQL, catalog tables, and pandas UDFs. - Use Databricks Runtime 17.3 LTS when your code depends on the package versions of that runtime. It also runs Spark.
- Use Python 3.13 or 3.14 for notebooks that need a newer Python and do not use Spark.
- Use Databricks Runtime 16.4 LTS only for existing code that depends on that runtime. It does not run Spark.
VS Code Server images do not run Spark. Use a JupyterLab image from the table for Spark work.
In an image without Spark support, starting a Spark session fails. On Python 3.13 and 3.14, the error says that pyspark is not installed. On Databricks Runtime 16.4 LTS, importing bsqai_spark fails because the module is not found.
Change the image for an existing dev environment
You cannot change the image of a running dev environment. To switch images:
- Stop the existing dev environment.
- Launch a new dev environment with the image you need.
- Mount the same workspace volume if you want to keep your files.