将数据存储管理升级到 SDK v2

Azure 机器学习数据存储将连接信息安全地存储在 Azure 上的数据存储中，因此无需在脚本中对其进行编码。 与 V1 相比，V2 数据存储概念大部分保持不变。 区别在于，我们不会通过 AzureML 数据存储支持类似 SQL 的数据源。 我们将通过 AzureML 数据导入和导出功能支持类似 SQL 的数据源。

本文比较 SDK v1 和 SDK v2 中的方案。

通过 account_key 从 Azure Blob 容器创建数据存储

• SDK v1

  blob_datastore_name='azblobsdk' # Name of the datastore to workspace
  container_name=os.getenv("BLOB_CONTAINER", "<my-container-name>") # Name of Azure blob container
  account_name=os.getenv("BLOB_ACCOUNTNAME", "<my-account-name>") # Storage account name
  account_key=os.getenv("BLOB_ACCOUNT_KEY", "<my-account-key>") # Storage account access key
  
  blob_datastore = Datastore.register_azure_blob_container(workspace=ws, 
                                                           datastore_name=blob_datastore_name, 
                                                           container_name=container_name, 
                                                           account_name=account_name,
                                                           account_key=account_key)
  

• SDK v2

  from azure.ai.ml.entities import AzureBlobDatastore
  from azure.ai.ml import MLClient
  
  ml_client = MLClient.from_config()
  
  store = AzureBlobDatastore(
      name="blob-protocol-example",
      description="Datastore pointing to a blob container using wasbs protocol.",
      account_name="mytestblobstore",
      container_name="data-container",
      protocol="wasbs",
      credentials={
          "account_key": "XXXxxxXXXxXXXXxxXXXXXxXXXXXxXxxXxXXXxXXXxXXxxxXXxxXXXxXxXXXxxXxxXXXXxxxxxXXxxxxxxXXXxXXX"
      },
  )
  
  ml_client.create_or_update(store)
  

通过 sas_token 从 Azure Blob 容器创建数据存储

• SDK v1

  blob_datastore_name='azblobsdk' # Name of the datastore to workspace
  container_name=os.getenv("BLOB_CONTAINER", "<my-container-name>") # Name of Azure blob container
  sas_token=os.getenv("BLOB_SAS_TOKEN", "<my-sas-token>") # Sas token
  
  blob_datastore = Datastore.register_azure_blob_container(workspace=ws, 
                                                           datastore_name=blob_datastore_name, 
                                                           container_name=container_name, 
                                                           sas_token=sas_token)
  

• SDK v2

  from azure.ai.ml.entities import AzureBlobDatastore
  from azure.ai.ml import MLClient
  
  ml_client = MLClient.from_config()
  
  store = AzureBlobDatastore(
      name="blob-sas-example",
      description="Datastore pointing to a blob container using SAS token.",
      account_name="mytestblobstore",
      container_name="data-container",
      credentials=SasTokenCredentials(
          sas_token= "?xx=XXXX-XX-XX&xx=xxxx&xxx=xxx&xx=xxxxxxxxxxx&xx=XXXX-XX-XXXXX:XX:XXX&xx=XXXX-XX-XXXXX:XX:XXX&xxx=xxxxx&xxx=XXxXXXxxxxxXXXXXXXxXxxxXXXXXxxXXXXXxXXXXxXXXxXXxXX"
      ),
  )
  
  ml_client.create_or_update(store)
  

通过基于标识的身份验证从 Azure Blob 容器创建数据存储

• SDK v1

blob_datastore = Datastore.register_azure_blob_container(workspace=ws,
                                                      datastore_name='credentialless_blob',
                                                      container_name='my_container_name',
                                                      account_name='my_account_name')

• SDK v2

  from azure.ai.ml.entities import AzureBlobDatastore
  from azure.ai.ml import MLClient
  
  ml_client = MLClient.from_config()
  
  store = AzureBlobDatastore(
      name="",
      description="",
      account_name="",
      container_name=""
  )
  
  ml_client.create_or_update(store)
  

从工作区获取数据存储

• SDK v1

  # Get a named datastore from the current workspace
  datastore = Datastore.get(ws, datastore_name='your datastore name')
  

  # List all datastores registered in the current workspace
  datastores = ws.datastores
  for name, datastore in datastores.items():
      print(name, datastore.datastore_type)
  

• SDK v2

  from azure.ai.ml import MLClient
  from azure.identity import DefaultAzureCredential
  
  #Enter details of your AzureML workspace
  subscription_id = '<SUBSCRIPTION_ID>'
  resource_group = '<RESOURCE_GROUP>'
  workspace_name = '<AZUREML_WORKSPACE_NAME>'
  
  ml_client = MLClient(credential=DefaultAzureCredential(),
                       subscription_id=subscription_id, 
                       resource_group_name=resource_group)
  
  datastore = ml_client.datastores.get(datastore_name='your datastore name')
  

SDK v1 和 SDK v2 中关键功能的映射

后续步骤

有关详细信息，请参阅：

• 创建数据存储
• 在作业中读取和写入数据
• V2 数据存储操作