如何在Vega中转换嵌套JSON数据以绘制Days-Value折线图？

问题背景 #

原始JSON数据格式：

{"groups":[{"days":[17,22,23,27,31],"experiment":"S1","value":[[19.77,23.92,22.19,20.47,25.85],[21.78,21.33,19.35,19.06,21.38],[18.86,24.47,19.40,20.50,24.13]]},{"days":[18,23,24,29,31],"experiment":"S2","value":[[12.77,24.92,27.19,21.47,29.85],[21.78,21.33,19.35,19.06,21.38],[18.86,24.47,19.40,20.50,24.13]]}, {"days":[27,32,33,37,41],"experiment":"S3","value":[[19.77,23.92,22.19,20.47,25.85],[21.78,21.33,19.35,19.06,21.38],[18.86,24.47,19.40,20.50,24.13]]}],"individuals":[]}

需求：将数据转换为包含experiment、day（原days数组拆分后）、value（原嵌套数组拆分后）的表格格式，用于绘制折线图。

现有数据加载代码：

"data": [
{
  "name": "table",
  "url": "path_to_json", 
  "transform": [
  ## 需要在此添加筛选groups并丢弃individuals的转换步骤 
  {
    "type": "flatten",
    "fields": ["value"],
    "as": ["value"]
  },
  {
    "type": "flatten",
    "fields": ["value", "days"],
    "as": ["value", "day"]
  }]
}

解决方案 #

在现有转换流程的最开头添加两个步骤，实现提取groups数据并丢弃无关字段：

完整转换代码如下：

"data": [
{
  "name": "table",
  "url": "path_to_json", 
  "transform": [
    {
      "type": "flatten",
      "fields": ["groups"],
      "as": ["group"]
    },
    {
      "type": "project",
      "fields": ["group.experiment", "group.days", "group.value"],
      "as": ["experiment", "days", "value"]
    },
    {
      "type": "flatten",
      "fields": ["value"],
      "as": ["value"]
    },
    {
      "type": "flatten",
      "fields": ["value", "days"],
      "as": ["value", "day"]
    }
  ]
}
]

步骤说明 #

• 第一步flatten：将原始数据顶层的groups数组展开，每个数组元素（单个实验组数据）转为独立条目，用group字段存储。
• 第二步project：从group字段中提取experiment、days、value属性并重命名为对应字段，自动丢弃原始数据的individuals字段和临时group字段。
• 后续两个flatten步骤：继续拆分嵌套的value数组，最终生成每行对应一组experiment、day、value的表格格式，满足折线图的数据要求。

内容的提问来源于stack exchange，提问作者Osman Mamun