pandas 将 json 转成 csv

今天分享的是【月小水长】pandas 三十六计系列的第八篇，一个小工具，将 json 文件转成 csv 文件。

文件格式是表，文件内容是里，只要里子一样，外表是可以像穿衣一样随便换的，就像在 MySQL 中，可以任意导入导出 SQL、csv、json 等文件一样。

假设我们有一个这样的 json 文件：

{
  "4893424946515214": {
    "mid": "4893424946515214",
    "weibo_link": "https://weibo.com/2803301701/MDcporkU6",
    "text": "据悉，全城月季花已逐渐进入盛花期。",
    "publish_time": "2023-04-22 20:34:45",
    "user_link": "https://weibo.com/u/2803301701",
    "user_name": "人民日报",
    "reposts_count": 55,
    "comments_count": 92,
    "attitudes_count": 298
  },
  "4893416880346795": {
    "mid": "4893416880346795",
    "weibo_link": "https://weibo.com/2803301701/MDcco1sdt",
    "text": "4月22日，陕西西安。游客发视频... ",
    "publish_time": "2023-04-22 20:02:42",
    "user_link": "https://weibo.com/u/2803301701",
    "user_name": "人民日报",
    "reposts_count": 119,
    "comments_count": 249,
    "attitudes_count": 785
  },
  "4893410513127118": {
    "mid": "4893410513127118",
    "weibo_link": "https://weibo.com/2803301701/MDc27d7vo",
    "text": "第54个世界地球日，江豚回家路还有多远...",
    "publish_time": "2023-04-22 19:37:24",
    "user_link": "https://weibo.com/u/2803301701",
    "user_name": "人民日报",
    "reposts_count": 119,
    "comments_count": 145,
    "attitudes_count": 463
  }
}

现在要转成下面这样的 csv：

mid	weibo_link	text	publish_time	user_link	user_name	reposts_count	comments_count	attitudes_count
4893424946515214	https://weibo.com/2803301701/MDcporkU6	据悉，全城月季花已逐渐进入盛花期。	2023-04-22 20:34:45	https://weibo.com/u/2803301701	人民日报	55	92	298
4893416880346795	https://weibo.com/2803301701/MDcco1sdt	4月22日，陕西西安。游客发视频…	2023-04-22 20:02:42	https://weibo.com/u/2803301701	人民日报	119	249	785
4893410513127118	https://weibo.com/2803301701/MDc27d7vo	第54个世界地球日，江豚回家路还有多远…	2023-04-22 19:37:24	https://weibo.com/u/2803301701	人民日报	119	145	463

只需要运行下面这份代码：

# -*- coding: utf-8 -*-
# author:           inspurer(月小水长)
# create_time:      2023/4/22 20:49
# 运行环境           Python3.6+
# github            https://github.com/inspurer
# website           https://buyixiao.github.io/
# 微信公众号         月小水长

import json
import pandas as pd


def convert_json_to_csv(input_json_path, output_csv_path):
    with open(input_json_path, mode='r', encoding='utf-8-sig') as f:
        input_json = json.loads(f.read())

    data_list = []

    data_cols = input_json[list(input_json.keys())[0]].keys()
    for a_weibo in input_json.values():
        data_list.append(list(a_weibo.values()))
    df = pd.DataFrame(data_list, columns=data_cols)

    df.to_csv(output_csv_path, index=False, encoding='utf-8-sig')


convert_json_to_csv('./data/2803301701.json', './data/2803301701.csv')