使用JQ扁平化数组键时出现笛卡尔积问题求助

原始JSON数据：

{
  "name": "vehicles",
  "details": [
    {
      "type": "passenger",
      "engine": {
        "type": "electric"
      }
    },
    {
      "type": "commercial",
      "engine": {
        "type": "diesel"
      }
    }
  ]
}

需求是将每个details元素中的engine.type字段重命名为engineType，同时删除原有的engine.type字段。

尝试执行的错误命令：

echo $json|jq '(.details[].engineType = .details[].engine.type)|del(.details[].engine.type)'

得到的错误结果（笛卡尔积输出）：

{
  "name": "vehicles",
  "details": [
    {
      "type": "passenger",
      "engine": {},
      "engineType": "electric"
    },
    {
      "type": "commercial",
      "engine": {},
      "engineType": "electric"
    }
  ]
}
{
  "name": "vehicles",
  "details": [
    {
      "type": "passenger",
      "engine": {},
      "engineType": "diesel"
    },
    {
      "type": "commercial",
      "engine": {},
      "engineType": "diesel"
    }
  ]
}

解决方法 #

问题核心在于原始命令中，.details[].engineType = .details[].engine.type会遍历所有details元素的engine.type值，并将每个值赋值给所有details元素的engineType字段，从而产生笛卡尔积。正确的做法是对每个details元素单独处理：

方法1：使用|=操作符逐个处理数组元素 #

echo "$json" | jq '.details[] |= (.engineType = .engine.type | del(.engine.type))'

方法2：重新构造details数组 #

echo "$json" | jq '.details = [.details[] | .engineType = .engine.type | del(.engine.type)]'

执行任一命令后，都会得到正确输出：

{
  "name": "vehicles",
  "details": [
    {
      "type": "passenger",
      "engine": {},
      "engineType": "electric"
    },
    {
      "type": "commercial",
      "engine": {},
      "engineType": "diesel"
    }
  ]
}

如果希望连空的engine对象也一起删除，可以调整命令为：

echo "$json" | jq '.details[] |= (.engineType = .engine.type | del(.engine))'

输出结果：

{
  "name": "vehicles",
  "details": [
    {
      "type": "passenger",
      "engineType": "electric"
    },
    {
      "type": "commercial",
      "engineType": "diesel"
    }
  ]
}

内容的提问来源于stack exchange，提问作者Leigh K