Datasource ReferenceΒΆ

To have a Datasource produce Data Assets of a custom type, such as when adding custom expectations by subclassing an existing DataAsset type, use the data_asset_type parameter to configure the datasource to load and return DataAssets of the custom type.

For example:

datasources:
  pandas:
    class_name: PandasDatasource
    data_asset_type:
      class_name: MyCustomPandasAsset
      module_name: internal_pandas_assets

Given the above configuration, we can observe the following:

>>> batch_kwargs = {
...   "datasource": "pandas",
...   "dataset": {"a": [1, 2, 3]}
... }
>>> batch = context.get_batch(batch_kwargs, my_suite)
>>> isinstance(batch, MyCustomPandasAsset)
True

Last updated: Aug 13, 2020