我有一个数据格式的数据集,看起来像这样:
[{'session_id': ['X061RFWB06K9V'],
'unix_timestamp': [1442503708],
'cities': ['New York NY, Newark NJ'],
'user': [[{'user_id': 2024,
'joining_date': '2015-03-22',
'country': 'UK'}]]},
{'session_id': ['5AZ2X2A9BHH5U'],
'unix_timestamp': [1441353991],
'cities': ['New York NY, Jersey City NJ, Philadelphia PA'],
'user': [[{'user_id': 2853,
'joining_date': '2015-03-28',
'country': 'DE'}]]},
{'session_id': ['SHTB4IYAX4PX6'],
'unix_timestamp': [1440843490],
'cities': ['San Antonio TX'],
'user': [[{'user_id': 10958,
'joining_date': '2015-03-06',
'country': 'UK'}]]}
我正在使用熊猫并对其进行处理,当我使用read_json时,会得到以下信息:
cities session_id unix_timestamp user
0 [New York NY, Newark NJ] [X061RFWB06K9V] [1442503708] [[{'user_id': 2024, 'joining_date': '2015-03-2...
1 [New York NY, Jersey City NJ, Philadelphia PA] [5AZ2X2A9BHH5U] [1441353991] [[{'user_id': 2853, 'joining_date': '2015-03-2...
2 [San Antonio TX] [SHTB4IYAX4PX6] [1440843490] [[{'user_id': 10958, 'joining_date': '2015-03-...
我该如何处理这些数据以使其具有更好的格式? 这是数据定义:
列:
session_id
:会话ID。 unix_timestamp
:会话开始时间的unix时间戳cities
:在同一会话中搜索的唯一城市user
:
user_id
:用户的ID joining_date
:用户创建帐户时country
:用户所在的地方我尝试使用json_normalize,但始终收到错误消息:
AttributeError:“ int”对象没有属性“ values”
以及不同类型的错误。请帮助
答案 0 :(得分:1)
您可以使用将其完全展平的函数,然后重建数据框:
import re
import pandas as pd
import numpy as np
jsonData = [{'session_id': ['X061RFWB06K9V'],
'unix_timestamp': [1442503708],
'cities': ['New York NY, Newark NJ'],
'user': [[{'user_id': 2024,
'joining_date': '2015-03-22',
'country': 'UK'}]]},
{'session_id': ['5AZ2X2A9BHH5U'],
'unix_timestamp': [1441353991],
'cities': ['New York NY, Jersey City NJ, Philadelphia PA'],
'user': [[{'user_id': 2853,
'joining_date': '2015-03-28',
'country': 'DE'}]]},
{'session_id': ['SHTB4IYAX4PX6'],
'unix_timestamp': [1440843490],
'cities': ['San Antonio TX'],
'user': [[{'user_id': 10958,
'joining_date': '2015-03-06',
'country': 'UK'}]]} ]
def flatten_json(y):
out = {}
def flatten(x, name=''):
if type(x) is dict:
for a in x:
flatten(x[a], name + a + '_')
elif type(x) is list:
i = 0
for a in x:
flatten(a, name + str(i) + '_')
i += 1
else:
out[name[:-1]] = x
flatten(y)
return out
flat = flatten_json(jsonData)
results = pd.DataFrame()
columns_list = list(flat.keys())
for item in columns_list:
row_idx = re.findall(r'(\d+)\_', item )[0]
column = item.replace(row_idx+'_', '',1)
column = column.replace('_0', '')
row_idx = int(row_idx)
value = flat[item]
results.loc[row_idx, column] = value
# If you don't want to expand/split the `cities` column, remove line below
results = results.join(results['cities'].str.split(',', expand=True).add_prefix('cities_').fillna(np.nan))
print (results)
输出:
print (results.to_string())
session_id unix_timestamp cities user_user_id user_joining_date user_country cities_0 cities_1 cities_2
0 X061RFWB06K9V 1.442504e+09 New York NY, Newark NJ 2024.0 2015-03-22 UK New York NY Newark NJ NaN
1 5AZ2X2A9BHH5U 1.441354e+09 New York NY, Jersey City NJ, Philadelphia PA 2853.0 2015-03-28 DE New York NY Jersey City NJ Philadelphia PA
2 SHTB4IYAX4PX6 1.440843e+09 San Antonio TX 10958.0 2015-03-06 UK San Antonio TX NaN NaN