What is the process of programmatically generating data asset using Python code from a data store URI in Azure ML studio?

edit

The process of programmatically generating data asset using Python code from a data store URI in Azure ML studio involves the following steps:

1. Import the necessary packages:

   from azureml.core import Workspace, Datastore, Dataset
   import pandas as pd
   

2. Initialize a Workspace object using the Azure ML SDK:

   ws = Workspace.from_config()
   

3. Access the Datastore object using the data store URI:

   datastore = Datastore.get(ws, datastore_name='<datastore_name>')
   

4. Read data from the Datastore using the Dataset object:

   dataset = Dataset.Tabular.from_delimited_files(path=(datastore, '<dataset_file_path>'))
   df = dataset.to_pandas_dataframe()
   

5. Manipulate the data as needed using pandas DataFrame operations:

   df = df.drop(columns=['column_to_drop'])
   df['new_column'] = df['column_a'] + df['column_b']
   

6. Write the updated data back to the Datastore using the Dataset object:

   updated_dataset = Dataset.Tabular.register_pandas_dataframe(df, target=(datastore, '<updated_dataset_file_path>'), name='<dataset_name>', description='<description>')
   

7. The updated Dataset can be used further by passing it as input to the Azure ML training pipeline or other workflows.