将json数据集格式转换成csv格式

1、引入相关库

import numpy as np
import json
import csv
import pandas as pd

2、从json数据集里获取数据

with open('E:/database/Amson/Toys/Toys_and_Games_5.json','r')as f:
    data = f.readlines()

# 解析 JSON 数据
json_data = [json.loads(line) for line in data]

# 现在，json_data 是一个包含所有 JSON 对象的列表
# 可以按照需要访问和处理这些对象
click_v = []
click_list = []
user_ids = set()
item_ids = set()
user_count = 0
item_count = 0 
rate_count = 0
for obj in json_data:
    uid = obj['reviewerID']
    iid = obj['asin']
    rating = obj['overall']
    review = obj.get('reviewText','')
    user_ids.add(uid)
    item_ids.add(iid)
    if rating > rate_count:
        rate_count = rating
    click_v.append([uid,iid,rating,review])
    click_list.append([uid,iid,rating])
# 打印提取的数据
user_count = len(user_ids)
item_count = len(item_ids)
print(user_count,item_count,rate_count)
#print(click_v)
print(click_list[:10])

输出结果：

3、创建用户ID和物品ID的映射字典，将字符串ID替换为数字索引

# 创建用户ID和物品ID的映射字典
user_id_map = {uid: i for i, uid in enumerate(set(row[0] for row in click_v))}
item_id_map = {iid: i for i, iid in enumerate(set(row[1] for row in click_v))}

# 将字符串ID替换为数字索引
for row in click_v:
    row[0] = user_id_map[row[0]]
    row[1] = item_id_map[row[1]]

print(click_v[:10])  

结果：

4、将数据存入csv文件里

file_name = 'E:/database/Amson/Toys/Toys_and_Games_5.csv'
header = ['uid','iid','rating','review']
# 使用 csv 模块写入数据到 CSV 文件
with open(file_name, 'w', newline='', encoding='utf-8') as f:
    writer = csv.writer(f)
    # 写入表头
    writer.writerow(header)   
    # 写入数据
    writer.writerows(click_v)

5、查看该csv文件里的内容

data = pd.read_csv('E:/database/Amson/Toys/Toys_and_Games_5.csv')
data.head()

结果：