共用方式為


Microsoft.MachineLearningServices 工作區/數據集 2020-05-01-preview

Bicep 資源定義

工作區/資料集資源類型可以使用目標作業來部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/datasets 資源,請將下列 Bicep 新增至範本。

resource symbolicname 'Microsoft.MachineLearningServices/workspaces/datasets@2020-05-01-preview' = {
  parent: resourceSymbolicName
  datasetType: 'string'
  name: 'string'
  parameters: {
    header: 'string'
    includePath: bool
    partitionFormat: 'string'
    path: {
      dataPath: {
        datastoreName: 'string'
        relativePath: 'string'
      }
      httpUrl: 'string'
    }
    query: {
      datastoreName: 'string'
      query: 'string'
    }
    separator: 'string'
    sourceType: 'string'
  }
  registration: {
    description: 'string'
    name: 'string'
    tags: {
      {customized property}: 'string'
    }
  }
  skipValidation: bool
  timeSeries: {
    coarseGrainTimestamp: 'string'
    fineGrainTimestamp: 'string'
  }
}

屬性值

DatasetCreateRequestParameters

名稱 說明 價值
頁眉 標頭類型。 'all_files_have_same_headers'
'combine_all_files_headers'
'no_headers'
'only_first_file_has_headers'
includePath 布爾值,將路徑資訊保留為數據集中的數據行。 預設為 False。 這在讀取多個檔案時很有用,而且想要知道特定記錄的來源檔案,或保留檔案路徑中有用的資訊。 布爾 (bool)
partitionFormat 每個路徑的數據分割信息都會根據指定的格式擷取到數據行中。 格式元件 '{column_name}' 會建立字符串數據行,而 '{column_name:yy/MM/dd/HH/mm/ss}' 會建立 datetime 數據行,其中 'yyyy'、'MM'、'dd'、'HH'、'mm' 和 'ss' 用來擷取日期時間類型的年、月、日、小時、分和秒。 格式應該從第一個數據分割索引鍵的位置開始,直到檔案路徑結尾為止。 例如,假設路徑為 『.。/USA/2019/01/01/data.parquet',其中分割區依國家/地區和時間, partition_format='/{CountryOrRegion}/{PartitionDate:yyyy/MM/dd}/data.csv' 會建立值為 'USA' 的字符串數據行 'CountryOrRegion',以及值為 '2019-01-01 的日期時間數據行 'PartitionDate' 字串
路徑 DatasetCreateRequestParametersPath
查詢 DatasetCreateRequestParametersQuery
分隔符號 用來分割 『delimited_files』 sourceType 資料行的分隔符。 字串
資源類型 數據源類型。 'delimited_files'
'json_lines_files'
'parquet_files'

DatasetCreateRequestParametersPath

名稱 說明 價值
dataPath DatasetCreateRequestParametersPathDataPath
httpUrl Http URL。 字串

DatasetCreateRequestParametersPathDataPath

名稱 說明 價值
datastoreName 數據存放區名稱。 字串
relativePath 數據存放區內的路徑。 字串

DatasetCreateRequestParametersQuery

名稱 說明 價值
datastoreName SQL/PostgreSQL/MySQL 資料存放區名稱。 字串
查詢 SQL Quey。 字串

DatasetCreateRequestRegistration

名稱 說明 價值
說明 數據集的描述。 字串
名稱 數據集的名稱。 字串
標記 與數據集相關聯的標記。 DatasetCreateRequestRegistrationTags

DatasetCreateRequestRegistrationTags

名稱 說明 價值

DatasetCreateRequestTimeSeries

名稱 說明 價值
coarseGrainTimestamp 要當做 CoarseGrainTimestamp 使用的數據行名稱。 只有在指定 'fineGrainTimestamp' 且不能與 'fineGrainTimestamp' 相同時,才能使用。 字串
fineGrainTimestamp 要當做 FineGrainTimestamp 使用的數據行名稱 字串

Microsoft.MachineLearningServices/workspaces/datasets

名稱 說明 價值
datasetType 指定數據集類型。 'file'
表格式( 必要 )
名稱 資源名稱 字串 (必要)
參數 DatasetCreateRequestParameters (必要)
家長 在 Bicep 中,您可以指定子資源的父資源。 只有在父資源外部宣告子資源時,才需要新增這個屬性。

如需詳細資訊,請參閱 父資源外部的子資源
類型的資源符號名稱:工作區
註冊 DatasetCreateRequestRegistration (必要)
skipValidation 略過驗證,以確保在註冊之前可以從數據集載入數據。 布爾 (bool)
timeSeries DatasetCreateRequestTimeSeries

ARM 樣本資源定義

工作區/資料集資源類型可以使用目標作業來部署:

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/datasets 資源,請將下列 JSON 新增至範本。

{
  "type": "Microsoft.MachineLearningServices/workspaces/datasets",
  "apiVersion": "2020-05-01-preview",
  "name": "string",
  "datasetType": "string",
  "parameters": {
    "header": "string",
    "includePath": "bool",
    "partitionFormat": "string",
    "path": {
      "dataPath": {
        "datastoreName": "string",
        "relativePath": "string"
      },
      "httpUrl": "string"
    },
    "query": {
      "datastoreName": "string",
      "query": "string"
    },
    "separator": "string",
    "sourceType": "string"
  },
  "registration": {
    "description": "string",
    "name": "string",
    "tags": {
      "{customized property}": "string"
    }
  },
  "skipValidation": "bool",
  "timeSeries": {
    "coarseGrainTimestamp": "string",
    "fineGrainTimestamp": "string"
  }
}

屬性值

DatasetCreateRequestParameters

名稱 說明 價值
頁眉 標頭類型。 'all_files_have_same_headers'
'combine_all_files_headers'
'no_headers'
'only_first_file_has_headers'
includePath 布爾值,將路徑資訊保留為數據集中的數據行。 預設為 False。 這在讀取多個檔案時很有用,而且想要知道特定記錄的來源檔案,或保留檔案路徑中有用的資訊。 布爾 (bool)
partitionFormat 每個路徑的數據分割信息都會根據指定的格式擷取到數據行中。 格式元件 '{column_name}' 會建立字符串數據行,而 '{column_name:yy/MM/dd/HH/mm/ss}' 會建立 datetime 數據行,其中 'yyyy'、'MM'、'dd'、'HH'、'mm' 和 'ss' 用來擷取日期時間類型的年、月、日、小時、分和秒。 格式應該從第一個數據分割索引鍵的位置開始,直到檔案路徑結尾為止。 例如,假設路徑為 『.。/USA/2019/01/01/data.parquet',其中分割區依國家/地區和時間, partition_format='/{CountryOrRegion}/{PartitionDate:yyyy/MM/dd}/data.csv' 會建立值為 'USA' 的字符串數據行 'CountryOrRegion',以及值為 '2019-01-01 的日期時間數據行 'PartitionDate' 字串
路徑 DatasetCreateRequestParametersPath
查詢 DatasetCreateRequestParametersQuery
分隔符號 用來分割 『delimited_files』 sourceType 資料行的分隔符。 字串
資源類型 數據源類型。 'delimited_files'
'json_lines_files'
'parquet_files'

DatasetCreateRequestParametersPath

名稱 說明 價值
dataPath DatasetCreateRequestParametersPathDataPath
httpUrl Http URL。 字串

DatasetCreateRequestParametersPathDataPath

名稱 說明 價值
datastoreName 數據存放區名稱。 字串
relativePath 數據存放區內的路徑。 字串

DatasetCreateRequestParametersQuery

名稱 說明 價值
datastoreName SQL/PostgreSQL/MySQL 資料存放區名稱。 字串
查詢 SQL Quey。 字串

DatasetCreateRequestRegistration

名稱 說明 價值
說明 數據集的描述。 字串
名稱 數據集的名稱。 字串
標記 與數據集相關聯的標記。 DatasetCreateRequestRegistrationTags

DatasetCreateRequestRegistrationTags

名稱 說明 價值

DatasetCreateRequestTimeSeries

名稱 說明 價值
coarseGrainTimestamp 要當做 CoarseGrainTimestamp 使用的數據行名稱。 只有在指定 'fineGrainTimestamp' 且不能與 'fineGrainTimestamp' 相同時,才能使用。 字串
fineGrainTimestamp 要當做 FineGrainTimestamp 使用的數據行名稱 字串

Microsoft.MachineLearningServices/workspaces/datasets

名稱 說明 價值
apiVersion API 版本 '2020-05-01-preview'
datasetType 指定數據集類型。 'file'
表格式( 必要 )
名稱 資源名稱 字串 (必要)
參數 DatasetCreateRequestParameters (必要)
註冊 DatasetCreateRequestRegistration (必要)
skipValidation 略過驗證,以確保在註冊之前可以從數據集載入數據。 布爾 (bool)
timeSeries DatasetCreateRequestTimeSeries
類型 資源類型 'Microsoft.MachineLearningServices/workspaces/datasets'

使用範例

Azure 快速入門範本

下列 Azure 快速入門範本 部署此資源類型。

範本 說明
使用多個數據集建立 AML 工作區 & 資料存放區

部署至 Azure
此範本會建立具有多個數據集的 Azure Machine Learning 工作區,& 數據存放區。
從數據存放區 中的相對路徑建立檔案數據集

部署至 Azure
此範本會從 Azure Machine Learning 工作區中數據存放區中的相對路徑建立檔案數據集。
從 Web URL 在 AML 工作區中建立檔案數據集

部署至 Azure
此範本會從 Azure Machine Learning 工作區中的 Web URL 建立檔案數據集。
從數據存放區中的相對路徑建立表格式數據集

部署至 Azure
此範本會從 Azure Machine Learning 工作區中資料存放區中的相對路徑建立表格式數據集。
從 SQL/PostgreSQL/MySQL 數據存放區建立表格式數據集

部署至 Azure
此範本會在 Azure Machine Learning 工作區的 SQL/PostgreSQL/MySQL 數據存放區中,從 SQL 查詢建立表格式數據集。
從 Web URL 在 AML 工作區中建立表格式數據集

部署至 Azure
此範本會從 Azure Machine Learning 工作區中的 Web URL 建立表格式數據集。

Terraform (AzAPI 提供者) 資源定義

工作區/資料集資源類型可以使用目標作業來部署:

  • 資源群組

如需每個 API 版本中已變更屬性的清單,請參閱 變更記錄檔

資源格式

若要建立 Microsoft.MachineLearningServices/workspaces/datasets 資源,請將下列 Terraform 新增至範本。

resource "azapi_resource" "symbolicname" {
  type = "Microsoft.MachineLearningServices/workspaces/datasets@2020-05-01-preview"
  name = "string"
  body = {
    datasetType = "string"
    parameters = {
      header = "string"
      includePath = bool
      partitionFormat = "string"
      path = {
        dataPath = {
          datastoreName = "string"
          relativePath = "string"
        }
        httpUrl = "string"
      }
      query = {
        datastoreName = "string"
        query = "string"
      }
      separator = "string"
      sourceType = "string"
    }
    registration = {
      description = "string"
      name = "string"
      tags = {
        {customized property} = "string"
      }
    }
    skipValidation = bool
    timeSeries = {
      coarseGrainTimestamp = "string"
      fineGrainTimestamp = "string"
    }
  }
}

屬性值

DatasetCreateRequestParameters

名稱 說明 價值
頁眉 標頭類型。 'all_files_have_same_headers'
'combine_all_files_headers'
'no_headers'
'only_first_file_has_headers'
includePath 布爾值,將路徑資訊保留為數據集中的數據行。 預設為 False。 這在讀取多個檔案時很有用,而且想要知道特定記錄的來源檔案,或保留檔案路徑中有用的資訊。 布爾 (bool)
partitionFormat 每個路徑的數據分割信息都會根據指定的格式擷取到數據行中。 格式元件 '{column_name}' 會建立字符串數據行,而 '{column_name:yy/MM/dd/HH/mm/ss}' 會建立 datetime 數據行,其中 'yyyy'、'MM'、'dd'、'HH'、'mm' 和 'ss' 用來擷取日期時間類型的年、月、日、小時、分和秒。 格式應該從第一個數據分割索引鍵的位置開始,直到檔案路徑結尾為止。 例如,假設路徑為 『.。/USA/2019/01/01/data.parquet',其中分割區依國家/地區和時間, partition_format='/{CountryOrRegion}/{PartitionDate:yyyy/MM/dd}/data.csv' 會建立值為 'USA' 的字符串數據行 'CountryOrRegion',以及值為 '2019-01-01 的日期時間數據行 'PartitionDate' 字串
路徑 DatasetCreateRequestParametersPath
查詢 DatasetCreateRequestParametersQuery
分隔符號 用來分割 『delimited_files』 sourceType 資料行的分隔符。 字串
資源類型 數據源類型。 'delimited_files'
'json_lines_files'
'parquet_files'

DatasetCreateRequestParametersPath

名稱 說明 價值
dataPath DatasetCreateRequestParametersPathDataPath
httpUrl Http URL。 字串

DatasetCreateRequestParametersPathDataPath

名稱 說明 價值
datastoreName 數據存放區名稱。 字串
relativePath 數據存放區內的路徑。 字串

DatasetCreateRequestParametersQuery

名稱 說明 價值
datastoreName SQL/PostgreSQL/MySQL 資料存放區名稱。 字串
查詢 SQL Quey。 字串

DatasetCreateRequestRegistration

名稱 說明 價值
說明 數據集的描述。 字串
名稱 數據集的名稱。 字串
標記 與數據集相關聯的標記。 DatasetCreateRequestRegistrationTags

DatasetCreateRequestRegistrationTags

名稱 說明 價值

DatasetCreateRequestTimeSeries

名稱 說明 價值
coarseGrainTimestamp 要當做 CoarseGrainTimestamp 使用的數據行名稱。 只有在指定 'fineGrainTimestamp' 且不能與 'fineGrainTimestamp' 相同時,才能使用。 字串
fineGrainTimestamp 要當做 FineGrainTimestamp 使用的數據行名稱 字串

Microsoft.MachineLearningServices/workspaces/datasets

名稱 說明 價值
datasetType 指定數據集類型。 'file'
表格式( 必要 )
名稱 資源名稱 字串 (必要)
參數 DatasetCreateRequestParameters (必要)
parent_id 此資源為父系之資源的標識碼。 類型資源的標識碼:工作區
註冊 DatasetCreateRequestRegistration (必要)
skipValidation 略過驗證,以確保在註冊之前可以從數據集載入數據。 布爾 (bool)
timeSeries DatasetCreateRequestTimeSeries
類型 資源類型 “Microsoft.MachineLearningServices/workspaces/datasets@2020-05-01-preview”