Hi, for some reason i cant name a dataset type

spark.SparkDatasetV2

, it has to be

kedro_datasets.spark.SparkDatasetV2

. Im playing with 1.1.1, and databricks-connect Just to know if something has change, bc the docs say its the first way.

Hi @Emilio Vega How are you installing kedro-datasets? Can you copy paste the dependencies section of your pyproject.toml?

dependencies = [
    "ipython>=8.10",
    "jupyterlab>=3.0",
    "notebook",
    "kedro[jupyter]~=1.1.1",
    "kedro-viz>=6.7.0",
    "scikit-learn~=1.5.1",
    "seaborn~=0.12.1",
    "setuptools; python_version >= '3.12'",
    "databricks-connect~=16.0",
    "kedro-datasets>=9.1.1",
    "s3fs>=2025.12.0",
]

Thank you @Guillaume Tauzin

From what I gathered from the Kedro code, it tries to import it from

kedro_datasets

by adding it as a prefix and just trying it to see if it is there: https://github.com/kedro-org/kedro/blob/main/kedro/io/core.py#L552-L578 So there should not be a need to add

kedro_datasets.

at all. A bit of a guess here, but typically, you install kedro-datasets along with the extra corresponding to the dataset that you want to use in dependencies. So I would try to add the spark-related extra you need. In your case, it could be something like: dependencies = [ "ipython>=8.10", "jupyterlab>=3.0", "notebook", "kedro[jupyter]~=1.1.1", "kedro-viz>=6.7.0", "scikit-learn~=1.5.1", "seaborn~=0.12.1", "setuptools; python_version >= '3.12'", "databricks-connect~=16.0", "kedro-datasets[spark-all]>=9.1.1", # This changed "s3fs>=2025.12.0", ] I put

spark-all

just to be sure you have everything but in practice, it's best to select the specific one you need so that you don't install a massive amount of unused dependencies. You can see all the one available here: https://github.com/kedro-org/kedro-plugins/blob/main/kedro-datasets/pyproject.toml#L30-L61

Wow thats interesting! Its working now, at first i was trying to use extras, maybe something broke. All is running at Databricks so i asume it wont need extras. Thank you!