编写JOLT转换规范：将VT对象拆分为独立实体并关联元数据

问题原因 #

你当前的输出会把同一ChannelData下的VT数据合并成数组，核心原因是shift操作中使用了[&4]作为目标对象的索引：

• &4指向的是ChannelData数组的下标（从VT的*层级往上数，第4层是ChannelData数组的索引位）
• 同一个ChannelData下的所有VT元素都会被映射到同一个[&4]对象中，JOLT会自动将重复字段的多个值合并为数组。

正确的JOLT转换规范 #

要让每个VT对象成为独立的顶层对象，需要将目标数组的索引改为[]（自动为每个VT元素创建新的数组项），同时正确引用上层的字段：

[
  {
    "operation": "shift",
    "spec": {
      "StationDataList": {
        "StationData": {
          "ChannelData": {
            "*": {
              "Values": {
                "VT": {
                  "*": {
                    "@(4,stationId)": "[].stationId",
                    "@(4,timezone)": "[].timezone",
                    "@(4,name)": "[].name",
                    "@(1,channelId)": "[].channelId",
                    "t": "[].t",
                    "content": "[].content"
                  }
                }
              }
            }
          }
        }
      }
    }
  }
]

字段引用说明 #

• @(4,stationId)：从当前VT的*层级往上数4层，定位到StationData节点，获取stationId值
• @(1,channelId)：从当前VT的*层级往上数1层，定位到当前ChannelData节点，获取channelId值
• []：JOLT会自动为每个VT元素生成一个新的顶层数组对象，将对应字段逐一填充

预期输出 #

执行上述规范后，输出会是包含6个独立对象的数组（每个VT对应一个）：

[
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId1","t":"2023-12-13T00:15:00","content":-1},
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId1","t":"2023-12-13T00:30:00","content":-2},
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId1","t":"2023-12-13T00:45:00","content":-3},
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId2","t":"2023-12-13T01:00:00","content":1},
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId2","t":"2023-12-13T02:00:00","content":2.5},
  {"stationId":"123","timezone":"+01:00","name":"stationName","channelId":"channelId2","t":"2023-12-13T03:00:00","content":3}
]

内容的提问来源于stack exchange，提问作者Maxim Novakovski