如何在python中将JSON转换为Dataframe

Question

我有以下JSON格式，我需要将其转换为python中的数据帧。请让我知道，如何去做。

JSON：

 User Patterns
[{"Jane": [{"Thermostat": 20, "Days": [1, 2], "Hour": 6, "Minute": 43}],  
"John": [{"Thermostat": 18, "Days": [1, 2], "Hour": 0, "Minute": 15}], 
"Jen": [{"Thermostat": 22, "Days": [1, 2], "Hour": 10, "Minute": 1}]}]

我希望我的数据框看起来像：

User    Thermostat   Days     Hour   Minute
Jane    20           [1,2]    6      43  
John    18           [1,2]    0      15 
Jen     22           [1,2]    10     1

Answer 1

jstr = """[{"Jane": [{"Thermostat": 20, "Days": [1, 2], "Hour": 6, "Minute": 43}],  
"John": [{"Thermostat": 18, "Days": [1, 2], "Hour": 0, "Minute": 15}], 
"Jen": [{"Thermostat": 22, "Days": [1, 2], "Hour": 10, "Minute": 1}]}]"""

pd.DataFrame.from_dict(
    {k: v[0] for k, v in json.loads(jstr)[0].items()}, 'index'
).rename_axis('User').reset_index()

如果您想拆分Days列

df = pd.DataFrame.from_dict(
    {k: v[0] for k, v in json.loads(jstr)[0].items()}, 'index'
).rename_axis('User').reset_index()

df.drop('Days', 1).join(
    pd.DataFrame(df.Days.tolist()).rename(columns='Day{}'.format))

您可以将此功能化：

def read_my_json(jstr):
    return pd.DataFrame.from_dict(
        {k: v[0] for k, v in json.loads(jstr)[0].items()}, 'index'
    ).rename_axis('User').reset_index()