使用Newtonsoft.Json按JSON Path白名单裁剪JSON并保留原层级结构

核心逻辑 #

• 先把所有白名单JSON Path拆解为层级路径段，同时将每一级父路径也标记为合法，确保匹配节点的上层结构不会被误删
• 递归遍历原始JSON，仅保留落在合法路径范围内的节点，最终输出的结构和原始JSON的层级、键顺序完全一致
• 区别于普通JSON Path直接返回匹配值的逻辑，该方案会自动保留完整的父级路径结构

代码实现（Python，无第三方依赖，适配固定路径白名单） #

import json
from typing import Dict, List, Set

def split_json_path(path: str) -> List[str]:
    # 拆解 $.a.b.c 格式的JSON Path为路径段列表
    return path.lstrip("$.") .split(".")

def collect_allowed_prefixes(whitelist: List[str]) -> Set[tuple]:
    # 收集所有合法路径前缀（含完整匹配路径），用于判断节点是否需要保留
    allowed = set()
    for path_expr in whitelist:
        segments = split_json_path(path_expr)
        for i in range(1, len(segments) + 1):
            allowed.add(tuple(segments[:i]))
    return allowed

def filter_json_by_whitelist(source: Dict, allowed_prefixes: Set[tuple], current_path: tuple = ()) -> Dict:
    filtered_node = {}
    for key, value in source.items():
        new_path = current_path + (key,)
        if new_path not in allowed_prefixes:
            continue
        if isinstance(value, dict):
            child_node = filter_json_by_whitelist(value, allowed_prefixes, new_path)
            # 仅当子节点有有效内容时保留当前键，避免生成空对象
            if child_node:
                filtered_node[key] = child_node
        else:
            filtered_node[key] = value
    return filtered_node

# 测试运行
if __name__ == "__main__":
    # 输入原始JSON
    source_data = {
        "path1": {
            "path1Inner1": {"id": "id1"},
            "path1Inner2": {"id": "id2"}
        },
        "path2": {
            "path2Inner1": {"id": "id3"},
            "path2Inner2": {"id": "id4", "key": "key4"}
        }
    }
    # 输入白名单
    path_whitelist = [
        "$.path1.path1Inner1",
        "$.path2.path2Inner2.key"
    ]
    allowed_paths = collect_allowed_prefixes(path_whitelist)
    result = filter_json_by_whitelist(source_data, allowed_paths)
    # 输出结果
    print(json.dumps(result, indent=4, ensure_ascii=False))

输出结果 #

{
    "path1": {
        "path1Inner1": {
            "id": "id1"
        }
    },
    "path2": {
        "path2Inner2": {
            "key": "key4"
        }
    }
}

拓展说明 #

如果你的白名单需要支持通配符、数组索引、条件过滤等复杂JSON Path语法，可以引入jsonpath-ng库解析路径表达式，先找到所有匹配的节点全路径，再按上面的路径前缀匹配逻辑过滤即可。

内容的提问来源于stack exchange，提问作者ademchenko