如何在 Python 中创建 Azure 数据工厂管道活动之间的依赖关系

Question

在 Azure DataFactory 管道中，我尝试让两个 CopyActivities 按顺序运行，即第一个将数据从 blob 复制到 SQL 表，然后第二个将 SQL 表复制到另一个数据库。

我尝试了下面的代码，但结果管道没有建立在活动上的依赖性（从工作流图和 JSON 在 Azure UI 中检查）。 当我运行管道时，我收到如下错误消息： “ErrorResponseException：模板验证失败：模板操作‘我的第二个活动 nameScope’的‘runAfter’属性在‘1’行和‘22521’列包含不存在的行动。balababla……”

在 Azure UI 中手动添加依赖项后，我可以成功运行管道。

如果有人能指出示例代码 (Python/C#/Powershell) 或文档，我将不胜感激。 我的 Python 代码：

    def createDataFactoryRectStage(self,  
                               aPipelineName, aActivityStageName, aActivityAcquireName,
                               aRectFileName, aRectDSName,
                               aStageTableName, aStageDSName,
                               aAcquireTableName, aAcquireDSName):
    adf_client = self.__getAdfClient()

    ds_blob = AzureBlobDataset(linked_service_name = LinkedServiceReference(AZURE_DATAFACTORY_LS_BLOB_RECT), 
                               folder_path=PRJ_AZURE_BLOB_PATH_RECT, 
                               file_name = aRectFileName,
                               format = {"type": "TextFormat",
                                         "columnDelimiter": ",",
                                         "rowDelimiter": "",
                                         "nullValue": "\\N",
                                         "treatEmptyAsNull": "true",
                                         "firstRowAsHeader": "true",
                                         "quoteChar": "\"",})    
    adf_client.datasets.create_or_update(AZURE_RESOURCE_GROUP, AZURE_DATAFACTORY, aRectDSName, ds_blob)

    ds_stage= AzureSqlTableDataset(linked_service_name = LinkedServiceReference(AZURE_DATAFACTORY_LS_SQLDB_STAGE), 
                                   table_name='[dbo].[' + aStageTableName + ']')      
    adf_client.datasets.create_or_update(AZURE_RESOURCE_GROUP, AZURE_DATAFACTORY, aStageDSName, ds_stage)

    ca_blob_to_stage = CopyActivity(aActivityStageName, 
                                    inputs=[DatasetReference(aRectDSName)], 
                                    outputs=[DatasetReference(aStageDSName)], 
                                    source= BlobSource(), 
                                    sink= SqlSink(write_batch_size = AZURE_SQL_WRITE_BATCH_SIZE))

    ds_acquire= AzureSqlTableDataset(linked_service_name = LinkedServiceReference(AZURE_DATAFACTORY_LS_SQLDB_ACQUIRE), 
                                     table_name='[dbo].[' + aAcquireTableName + ']')      
    adf_client.datasets.create_or_update(AZURE_RESOURCE_GROUP, AZURE_DATAFACTORY, aAcquireDSName, ds_acquire)
    dep = ActivityDependency(ca_blob_to_stage, dependency_conditions =[DependencyCondition('Succeeded')])

    ca_stage_to_acquire = CopyActivity(aActivityAcquireName, 
                                       inputs=[DatasetReference(aStageDSName)], 
                                       outputs=[DatasetReference(aAcquireDSName)], 
                                       source= SqlSource(), 
                                       sink= SqlSink(write_batch_size = AZURE_SQL_WRITE_BATCH_SIZE),
                                       depends_on=[dep])

    p_obj = PipelineResource(activities=[ca_blob_to_stage, ca_stage_to_acquire], parameters={})

    return adf_client.pipelines.create_or_update(AZURE_RESOURCE_GROUP, AZURE_DATAFACTORY, aPipelineName, p_obj)

Answer 1

以防万一有人像我一样遇到与这个老问题相同的问题，python 代码中有一个微妙的错误

更改 dep 以使用活动名称，而不是对活动对象的引用使它对我有用。

dep = ActivityDependency(aActivityStageName, dependency_conditions =[DependencyCondition('Succeeded')])

Answer 2

这是C#中的一个示例，它基本上在管道内按顺序执行Chaining activities和链接活动。 请记住，在 ADFV1 中，我们必须将一个活动的输出配置为另一个活动的输入，以将它们链接起来并使它们相互依赖。

管道代码片段（注意 dependsOn 属性，它确保第二个活动在第一个活动成功运行后运行） -

static PipelineResource PipelineDefinition(DataFactoryManagementClient client) {
 Console.WriteLine("Creating pipeline " + pipelineName + "...");
 PipelineResource resource = new PipelineResource {
   Activities = new List < Activity > {
    new CopyActivity {
     Name = copyFromBlobToSQLActivity,
      Inputs = new List < DatasetReference > {
       new DatasetReference {
        ReferenceName = blobSourceDatasetName
       }
      },
      Outputs = new List<DatasetReference>
      {
       new DatasetReference {
        ReferenceName = sqlDatasetName
       }
      },
      Source = new BlobSource {},
      Sink = new SqlSink {}
    },
    new CopyActivity {
     Name = copyToSQLServerActivity,
      Inputs = new List < DatasetReference > {
       new DatasetReference {
        ReferenceName = sqlDatasetName
       }
      },
      Outputs = new List<DatasetReference>
      {
       new DatasetReference {
        ReferenceName = sqlDestinationDatasetName
       }
      },
      Source = new SqlSource {},
      Sink = new SqlSink {},
      DependsOn = new List < ActivityDependency > {
       new ActivityDependency {
        Activity = copyFromBlobToSQLActivity,
         DependencyConditions = new List < String > {
          "Succeeded"
         }
       }
      }
    }
   }
 };
 Console.WriteLine(SafeJsonConvert.SerializeObject(resource, client.SerializationSettings));
 return resource;
}

请在此处查看 ADFV2 教程以获得全面的解释和更多场景。