时间戳

首页 > 解决方案 > 如何在 Azure 数据工厂中处理数组中的项目/索引 - 数据流 - 展平

json - 如何在 Azure 数据工厂中处理数组中的项目/索引 - 数据流 - 展平

问题描述

我有一个如下所示的 JSON 源：

{
   "columnNames":[
      "Screw",
      "Type",
      "Class_Region",
      "Orders/Sales",
      "Quarter-3",
      "Previous Quarter %",
      "Previous Year %",
      "Quarter-2",
      "Previous Quarter %",
      "Previous Year %",
      "Quarter-1",
      "Previous Quarter %",
      "Previous Year %",
      "Quarter",
      "Previous Quarter %",
      "Previous Year %",
      "Total",
      "Previous Year %"
   ],
   "analysisData":[
      [
         "Single Screw",
         "Fields of Application (1.000€)",
         "Blown film",
         "Orders",
         "146.700",
         "119,4",
         "39,7",
         "147.088",
         "0,3",
         "17,0",
         "166.429",
         "13,1",
         "54,6",
         "119.562",
         "-28,2",
         "78,9",
         "579.779",
         "43,1"
      ],
      [
         "Single Screw",
         "Fields of Application (1.000€)",
         "Blown film",
         "Sales",
         "84.290",
         "-16,9",
         "-10,7",
         "122.121",
         "44,9",
         "-17,4",
         "102.930",
         "-15,7",
         "15,1",
         "128.227",
         "24,6",
         "26,4",
         "437.568",
         "1,0"
      ],
      [
         "Single Screw",
         "Fields of Application (1.000€)",
         "Flat film",
         "Orders",
         "56.077",
         "9,1",
         "13,8",
         "85.338",
         "52,2",
         "125,7",
         "41.544",
         "-51,3",
         "-8,5",
         "102.514",
         "146,8",
         "99,4",
         "285.473",
         "55,2"
      ]
   ],
   "resCode":0,
   "message":"OK!"
}

我想通过数据流将它加载到 Azure SQL 表中。我尝试了“Flatten”，我认为这是正确的步骤，但如何映射列？什么是正确的语法？

我需要的是这样的：

analysisData[0] -> Screw
analysisData[1] -> Type
analysisData[2] -> Class_Region

数组中索引的语法是什么（类似于 analysisData[0]）？

标签： jsonazure-data-factoryazure-data-factory-pipelineazure-data-flow

解决方案

而不是使用analysisData[]来调用值螺丝、类型、类区域。请使用columnNames[]以获得相同或取悦马克提到的评论，您必须将列名展平为行值，并且您必须再次使用 Pivot 将那些用于映射列。

推荐阅读