Reading dataset after uploading to storage

Question

hi,

I created various datasets but within my python notebook, how do i read it?

So currently if this is what I have:

x_train_df = pd.read_csv('data_reviews/x_train.csv')

what should I replace it with?

Answer 1

Hi, thanks for reaching out. I assume you registered a datastore and uploaded datasets to storage. To access the data from storage depends on whether it's structured or unstructured. For unstructured data you can access using FileDataset and for structured data you can access using TabularDataset. The following code snippet shows how to create dataset from datastore. For FileDataset, review Mount vs Download and sample notebook.

from azureml.core import Workspace, Datastore, Dataset  
  
datastore_name = 'your datastore name'  
  
# get existing workspace  
workspace = Workspace.from_config()  
      
# retrieve an existing datastore in the workspace by name  
datastore = Datastore.get(workspace, datastore_name)  
  
# create a TabularDataset from 3 file paths in datastore  
datastore_paths = [(datastore, 'weather/2018/11.csv'),  
                   (datastore, 'weather/2018/12.csv'),  
                   (datastore, 'weather/2019/*.csv')]  
  
weather_ds = Dataset.Tabular.from_delimited_files(path=datastore_paths)  

--please don't forget to Accept Answer if the reply is helpful. Thanks.--