如何使用json_normalize展平嵌套JSON，提取uuid与指定日期CardMarket价格？

你可以直接按路径手动提取所需字段，比调整json_normalize参数更高效，也不会出现大宽表的问题，完整代码如下：

import json
import pandas as pd

# 配置目标参数，可根据需求修改
TARGET_DATE = "2021-11-07"
PRICE_KEYS = ["paper", "cardmarket", "retail", "normal", TARGET_DATE]
RESULT_COL_NAME = f"paper.cardmarket.retail.normal.{TARGET_DATE}"

# 加载JSON数据
with open("AllPrices.json", "r", encoding="utf-8") as f:
    raw_data = json.load(f)
card_price_dict = raw_data["data"]

# 遍历提取所需字段
output_data = []
for uuid, price_info in card_price_dict.items():
    # 逐层取嵌套值，缺失则返回空
    current_val = price_info
    target_price = None
    try:
        for k in PRICE_KEYS:
            current_val = current_val[k]
        target_price = current_val
    except KeyError:
        pass
    output_data.append({
        "uuid": uuid,
        RESULT_COL_NAME: target_price
    })

# 转为DataFrame，可按需删除价格为空的行
df = pd.DataFrame(output_data)
# 可选：过滤无价格的记录
# df = df.dropna(subset=[RESULT_COL_NAME])

如果需要用json_normalize实现，可先将uuid转为字段再处理：

# 将UUID从字典键转为记录字段
processed_list = [{"uuid": k, **v} for k, v in card_price_dict.items()]
df = pd.json_normalize(processed_list)[["uuid", RESULT_COL_NAME]]

该方法会先展开所有嵌套字段，内存开销比手动提取高，仅适合数据量较小的场景。

内容的提问来源于stack exchange，提问作者fab