Jolt向map数组添加公共map元素避免同索引元素合并问题

问题场景 #

使用Jolt进行JSON转换时，输入JSON中geo数组的每个元素都包含元素为map类型的locations数组，需求为向所有locations数组的首位插入固定公共map元素，公共map内容如下：

{
  "city": "XYZ",
  "zip": "0000",
  "sublocations": [
    "included"
  ]
}

预期效果为公共map作为独立数组项，存在于每个locations数组的0索引位置。使用自行编写的shift类型Jolt spec执行转换后，0索引位置的公共map与原locations数组0索引位置的原有map发生同名字段合并，同名字段值被合并为数组类型，未得到预期的独立数组元素。

相关参考内容 #

输入JSON #

{
  "geo": [
    {
      "locationCode": "USA",
      "locations": [
        {
          "city": "AMS",
          "zip": "1234",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "EDF",
          "zip": "4567",
          "sublocations": [
            "included"
          ]
        }
      ]
    },
    {
      "locationCode": "FRN",
      "locations": [
        {
          "city": "CVF",
          "zip": "4321",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "GHJ",
          "zip": "8901",
          "sublocations": [
            "included"
          ]
        }
      ]
    }
  ]
}

预期输出JSON #

{
  "geo": [
    {
      "locationCode": "USA",
      "locations": [
        {
          "city": "XYZ",
          "zip": "0000",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "AMS",
          "zip": "1234",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "EDF",
          "zip": "4567",
          "sublocations": [
            "included"
          ]
        }
      ]
    },
    {
      "locationCode": "FRN",
      "locations": [
        {
          "city": "XYZ",
          "zip": "0000",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "CVF",
          "zip": "4321",
          "sublocations": [
            "included"
          ]
        },
        {
          "city": "GHJ",
          "zip": "8901",
          "sublocations": [
            "included"
          ]
        }
      ]
    }
  ]
}

原有错误Jolt Spec #

[{
  "operation": "shift",
  "spec": {
    "geo": {
      "*": {
        "locations": {
          "#XYZ": "geo[&2].locations[0].city",
          "#0000": "geo[&2].locations[0].zip",
          "#included": "geo[&2].locations[0].sublocations[]",
          "*": {
            "*": "geo[#4].locations[#2].&"
          }
        },
        "*": "geo[#2].&"
      }
    }
  }
}]

实际错误输出 #

{   "geo" : [ {
    "locationCode" : "USA",
    "locations" : [ {
      "city" : [ "XYZ", "AMS" ],
      "zip" : [ "0000", "1234" ],
      "sublocations" : [ "included", [ "included" ] ]
    }, {
      "city" : "EDF",
      "zip" : "4567",
      "sublocations" : [ "included" ]
    } ]   }, {
    "locationCode" : "FRN",
    "locations" : [ {
      "city" : [ "XYZ", "CVF" ],
      "zip" : [ "0000", "4321" ],
      "sublocations" : [ "included", [ "included" ] ]
    }, {
      "city" : "GHJ",
      "zip" : "8901",
      "sublocations" : [ "included" ]
    } ]   } ]
}

问题根因 #

原有spec存在两处逻辑错误：

• 遍历原locations数组元素时，直接使用#2获取原数组索引作为输出路径索引，原数组首个元素的索引为0，和硬编码写入公共元素的路径locations[0]完全重合。Jolt shift操作对相同路径下的同名字段会自动执行值合并，最终生成数组类型字段值，是公共元素和原首个元素被合并的核心原因。
• 公共元素的sublocations字段赋值时使用了[]通配符，和原数组中本身为数组类型的sublocations写入逻辑叠加后，会出现数组嵌套的异常结构。

修正方案 #

将原locations数组的所有元素输出索引整体+1，把0号位置预留给固定公共元素即可，修正后的可直接使用的spec如下：

[
  {
    "operation": "shift",
    "spec": {
      "geo": {
        "*": {
          "locations": {
            // 写入0位置的固定公共元素
            "#XYZ": "geo[&2].locations[0].city",
            "#0000": "geo[&2].locations[0].zip",
            "#included": "geo[&2].locations[0].sublocations[0]",
            // 原数组元素索引+1，从位置1开始写入，避免和公共元素位置冲突
            "*": {
              "*": "geo[#4].locations[#2+1].&"
            }
          },
          // 其余非locations字段正常透传
          "*": "geo[#2].&"
        }
      }
    }
  }
]

修正点说明 #

• 原locations数组元素的输出路径索引从#2调整为#2+1，原索引0的元素会写入位置1，原索引1的元素写入位置2，不会和位置0的公共元素路径冲突，从根本上避免字段合并问题。
• 公共元素的sublocations固定值直接写入[0]位置，解决数组嵌套问题。
• 其余字段透传逻辑保持不变，不影响原有其他字段的输出。

内容的提问来源于stack exchange，提问作者coder1234