Pandas 可以很方便的处理 JSON 数据

demo.json

[
    {
        "name":"张三",
        "age":23,
        "gender":true
    },
    {
        "name":"李四",
        "age":24,
        "gender":true
    },
    {
        "name":"王五",
        "age":25,
        "gender":false
    }
]

JSON 转换为 CSV
非常方便，只要通过 pd.read_json 读出JSON数据，再通过 df.to_csv 写入 CSV 即可

import pandas as pd

json_path = 'data/demo.json'

# 加载 JSON 数据
with open(json_path, 'r', encoding='utf8') as f:
    # 解析一个有效的JSON字符串并将其转换为Python字典
    df = pd.read_json(f.read())
    print(df.to_string())  # to_string() 用于返回 DataFrame 类型的数据，我们也可以直接处理 JSON 字符串。
    print('-' * 10)

    # 重新定义标题
    df.columns = ['姓名', '年龄', '性别']
    print(df)

    df.to_csv('data/result.csv', index=False, encoding='GB2312')

简单 JSON
从 URL 中读取 JSON 数据：

import pandas as pd
# 堆代码 duidaima.com
URL = 'https://static.runoob.com/download/sites.json'
df = pd.read_json(URL) # 和读文件一样
print(df)

输出：

     id    name             url  likes
0  A001    菜鸟教程  www.runoob.com     61
1  A002  Google  www.google.com    124
2  A003      淘宝  www.taobao.com     45

字典转化为 DataFrame 数据

import pandas as pd

s = {
    "col1": {"row1": 1, "row2": 2, "row3": 3},
    "col2": {"row1": "x", "row2": "y", "row4": "z"}
}

df = pd.DataFrame(s)
print(df)
print('-' * 10)

new_df = df.dropna()  # 数据清洗，删除包含空数据的行
print(new_df.to_string())
print('-' * 10)

df.fillna(99, inplace=True)  # fillna() 方法来替换一些空字段
print(df.to_string())

输出：不同的行会用 NaN 填充

      col1 col2
row1   1.0    x
row2   2.0    y
row3   3.0  NaN
row4   NaN    z
----------
      col1 col2
row1   1.0    x
row2   2.0    y
----------
      col1 col2
row1   1.0    x
row2   2.0    y
row3   3.0   99
row4  99.0    z

内嵌的 JSON 数据
nested_list.json 嵌套的JSON数据

{
  "school_name": "ABC primary school",
  "class": "Year 1",
  "students": [
    {
      "id": "A001",
      "name": "Tom",
      "math": 60,
      "physics": 66,
      "chemistry": 61
    },
    {
      "id": "A002",
      "name": "James",
      "math": 89,
      "physics": 76,
      "chemistry": 51
    },
    {
      "id": "A003",
      "name": "Jenny",
      "math": 79,
      "physics": 90,
      "chemistry": 78
    }
  ]
}

运行代码
data = json.loads(f.read()) 使用 Python JSON 模块载入数据。
json_normalize() 使用了参数 record_path 并设置为 ['students'] 用于展开内嵌的 JSON 数据 students。

import pandas as pd
import json

# 打印出结果JSON结构
with open('data/nested_list.json', 'r') as f:
    data = pd.read_json(f.read())
    print(data)

# 使用 Python JSON 模块载入数据
with open('data/nested_list.json', 'r') as f:
    data = json.loads(f.read())

# 展平数据-- json_normalize() 方法将内嵌的数据完整的解析出来：
df_nested_list = pd.json_normalize(data, record_path=['students'])
print(df_nested_list)
image

import pandas as pd
import json

data_path = 'data/nested_list.json'

print(('-' * 10) + ' 连同上级JSON值一起显示')
# 使用 Python JSON 模块载入数据
with open(data_path, 'r') as f:
    data = json.loads(f.read())

# 展平数据
df_nested_list = pd.json_normalize(
    data,
    record_path=['students'],
    meta=['school_name', 'class']
)
print(df_nested_list)

复杂 JSON
该数据嵌套了列表和字典，数据文件 nested_mix.json 如下
nested_mix.json

{
    "school_name": "local primary school",
    "class": "Year 1",
    "info": {
      "president": "John Kasich",
      "address": "ABC road, London, UK",
      "contacts": {
        "email": "admin@e.com",
        "tel": "123456789"
      }
    },
    "students": [
    {
        "id": "A001",
        "name": "Tom",
        "math": 60,
        "physics": 66,
        "chemistry": 61
    },
    {
        "id": "A002",
        "name": "James",
        "math": 89,
        "physics": 76,
        "chemistry": 51
    },
    {
        "id": "A003",
        "name": "Jenny",
        "math": 79,
        "physics": 90,
        "chemistry": 78
    }]
}

import pandas as pd
import json

# 使用 Python JSON 模块载入数据
with open('data/nested_mix.json', 'r') as f:
    data = json.loads(f.read())

df = pd.json_normalize(
    data,
    record_path=['students'],
    meta=[
        'class',
        ['info', 'president'],  # 类似 info.president
        ['info', 'contacts', 'tel']
    ]
)

print(df)
     id   name  math  ...   class  info.president info.contacts.tel
0  A001    Tom    60  ...  Year 1     John Kasich         123456789
1  A002  James    89  ...  Year 1     John Kasich         123456789
2  A003  Jenny    79  ...  Year 1     John Kasich         123456789

[3 rows x 8 columns]

读取内嵌数据中的一组数据
nested_deep.json

{
    "school_name": "local primary school",
    "class": "Year 1",
    "students": [
    {
        "id": "A001",
        "name": "Tom",
        "grade": {
            "math": 60,
            "physics": 66,
            "chemistry": 61
        }
 
    },
    {
        "id": "A002",
        "name": "James",
        "grade": {
            "math": 89,
            "physics": 76,
            "chemistry": 51
        }
       
    },
    {
        "id": "A003",
        "name": "Jenny",
        "grade": {
            "math": 79,
            "physics": 90,
            "chemistry": 78
        }
    }]
}

这里我们需要使用到 glom 模块来处理数据套嵌，glom 模块允许我们使用 . 来访问内嵌对象的属性。

第一次使用我们需要安装 glom：

pip3 install glom -i https://pypi.tuna.tsinghua.edu.cn/simple

import pandas as pd
from glom import glom

df = pd.read_json('nested_deep.json')

data = df['students'].apply(lambda row: glom(row, 'grade.math'))
print(data)
输出：

0    60
1    89
2    79