使用 Python 展开从 JSON 获取的字典

Question

此 JSON 结构来自 Spring Cloud 服务器配置：

{
  "item[0].subitem[0].key": "value1",
  "item[0].subitem[1].key": "value2",
  "item[1].subitem[0].key": "value3",
  "item[1].subitem[1].key": "value4"
}

我有一个 Flask 应用程序将使用该格式，但我需要将其转换为这种格式以进行进一步处理：

{
  "item": [
    {
      "subitem": [
        {
          "key": "value1"
        },
        {
          "key": "value1"
        }
      ]
    },
    {
      "subitem": [
        {
          "key": "value1"
        },
        {
          "key": "value1"
        }
      ]
    }
  ]
}

Answer 1

下面的代码显示了一个通用解决方案，其中包含比您的要求更多的测试用例：

import re
j = {
  "item[0].subitem[0].key": "value1",
  "item[0].subitem[1].key": "value2",
  "item[1].subitem[0].key": "value3",
  "item[1].subitem[1].key": "value4",
  "item2[0].subitem[0]": "value5",
  "item2[0].subitem[1]": "value6",
  "item2[1][0].key1": "value7",
  "item2[1][1].key2": "value8"
}
d = {}
for key, value in j.items():
    s = d
    tokens = re.findall(r'\w+', key)
    for count, (index, next_token) in enumerate(zip(tokens, tokens[1:] + [value]), 1):
        value = next_token if count == len(tokens) else [] if next_token.isdigit() else {}
        if isinstance(s, list):
            index = int(index)
            while index >= len(s):
                s.append(value)
        elif index not in s:
            s[index] = value
        s = s[index]

d

变为：

{'item': [{'subitem': [{'key': 'value1'}, {'key': 'value2'}]},
          {'subitem': [{'key': 'value3'}, {'key': 'value4'}]}],
 'item2': [{'subitem': ['value5', 'value6']},
           [{'key1': 'value7'}, {'key2': 'value8'}]]}

Answer 2

或使用为此制作的包：

https://github.com/amirziai/flatten

pip install flatten_json

支持展平和反展平。

Answer 3

我解释了为什么我不想在OP问题的评论中提供完整的解决方案。这应该足以暗示了：

def main():
    input = {
        "item[0].subitem[0].key": "value1",
        "item[0].subitem[1].key": "value2",
        "item[1].subitem[0].key": "value3",
        "item[1].subitem[1].key": "value4",
    }
    items = list(input.items())
    random.shuffle(items)
    shuffled = dict(items)

    result = normalize_keys(unflatten(shuffled))

    print(json.dumps(result, indent=2))

结果：

{
  "item": [
    {
      "subitem": [
        {
          "key": "value1"
        },
        {
          "key": "value2"
        }
      ]
    },
    {
      "subitem": [
        {
          "key": "value3"
        },
        {
          "key": "value4"
        }
      ]
    }
  ]
}