如何将CSV转JSON时的索引保持为整数类型？

场景与问题 #

我有一个包含法语和英语单词的CSV文件：

French,English
partie,part
histoire,history
chercher,search
seulement,only
police,police

尝试将其转换为JSON时，希望索引为整数类型，但生成的JSON中索引自动变成了字符串，用int()转换也无法解决。

我的代码：

import json
import pandas
data = pandas.read_csv("data/french_words.csv")
words = {int(index): {
    "French": row.French,
    "English": row.English,
    "known": None
    } for index, row in data.iterrows()
}
data.update(words)
with open("words.json", mode="w") as words_file:
    json.dump(words, words_file, indent=4)

print(words)

控制台输出（Python字典的键为整数）：

{
0: {'French': 'partie', 'English': 'part', 'known': None}, 
1: {'French': 'histoire', 'English': 'history', 'known': None}, 
2: {'French': 'chercher', 'English': 'search', 'known': None}, 
3: {'French': 'seulement', 'English': 'only', 'known': None}, 
4: {'French': 'police', 'English': 'police', 'known': None}, 
5: {'French': 'pensais', 'English': 'thought', 'known': None},
}  

生成的JSON文件输出（键变为字符串）：

{
    "0": {
        "French": "partie",
        "English": "part",
        "known": null
    },
    "1": {
        "French": "histoire",
        "English": "history",
        "known": null
    },
    "2": {
        "French": "chercher",
        "English": "search",
        "known": null
    },
    "3": {
        "French": "seulement",
        "English": "only",
        "known": null
    },
    "4": {
        "French": "police",
        "English": "police",
        "known": null
    },
    "5": {
        "French": "pensais",
        "English": "thought",
        "known": null
    },
}

原因 #

JSON语法明确规定，对象的键必须是字符串类型。即便Python字典中使用整数作为键，json.dump()在序列化时也会自动将整数键转换为字符串，这是JSON的固有特性，无法直接改变。

解决方案 #

方案1：改用数组（列表）结构（推荐） #

JSON数组的索引天然是整数，无需显式定义键，完美匹配需求。修改后的代码：

import json
import pandas

data = pandas.read_csv("data/french_words.csv")
# 将每行数据转为字典，组成列表
words_list = [
    {
        "French": row.French,
        "English": row.English,
        "known": None
    } for _, row in data.iterrows()
]

with open("words.json", mode="w") as words_file:
    json.dump(words_list, words_file, indent=4)

生成的JSON内容（数组结构，索引为整数）：

[
    {
        "French": "partie",
        "English": "part",
        "known": null
    },
    {
        "French": "histoire",
        "English": "history",
        "known": null
    },
    {
        "French": "chercher",
        "English": "search",
        "known": null
    },
    {
        "French": "seulement",
        "English": "only",
        "known": null
    },
    {
        "French": "police",
        "English": "police",
        "known": null
    },
    {
        "French": "pensais",
        "English": "thought",
        "known": null
    }
]

方案2：保留键值对结构，读取时转换键类型 #

如果必须保留键值对结构，只能接受JSON中的字符串键，后续读取JSON时再将键转换为整数：

# 读取JSON时转换键为整数
with open("words.json", mode="r") as words_file:
    json_data = json.load(words_file)
    words_dict = {int(key): value for key, value in json_data.items()}

内容的提问来源于stack exchange，提问作者Kbk12