Question

我有这个：

[['COMPANY:', [('U S News & World Report Inc', 63)]],
 ['ORGANIZATION:',
  [('Ashoka', 0),
   ('Innovators For The Public', 91),
   ('Us Environmental  Protection Agency', 55)]]]

我希望这能成为像熊猫一样的数据框：

NAME           ORGS                            PERCENT
Company        US News & World Report          63
Organization   Ashoka                          0
Organization   US Environmental Protection     55

Answer 1

data = [['COMPANY:', [('U S News & World Report Inc', 63)]],
 ['ORGANIZATION:',
  [('Ashoka', 0),
   ('Innovators For The Public', 91),
   ('Us Environmental  Protection Agency', 55)]]]

results = []

for name, rest in data:
    name = name.replace(":","").capitalize()
    for orgs, percent in rest:
        results.append( [name, orgs, percent] )

import pandas as pd

df = pd.DataFrame( results, columns=['NAME', 'ORGS', 'PERCENT'])

print df

结果：

           NAME                                 ORGS  PERCENT
0       Company          U S News & World Report Inc       63
1  Organization                               Ashoka        0
2  Organization            Innovators For The Public       91
3  Organization  Us Environmental  Protection Agency       55

Answer 2

这是一个你应该研究的from_dict读取方法，这种方法恰好适用于这种情况，只需要将列表转换为字典：

L=[['COMPANY:', [('U S News & World Report Inc', 63)]],
 ['ORGANIZATION:',
  [('Ashoka', 0),
   ('Innovators For The Public', 91),
   ('Us Environmental  Protection Agency', 55)]]]

In [160]:
df=pd.DataFrame.from_dict(dict(L), orient="index").stack().reset_index(level=0)
df['Name']=df[0].apply(lambda x: x[0])
df['Val']=df[0].apply(lambda x: x[1])
df['Type']=df.level_0.str.slice(stop=-1)
df.__delitem__(0)
df.__delitem__('level_0')

In [161]:
print df
                                  Name  Val          Type
0                               Ashoka    0  ORGANIZATION
1            Innovators For The Public   91  ORGANIZATION
2  Us Environmental  Protection Agency   55  ORGANIZATION
0          U S News & World Report Inc   63       COMPANY

将具有可变长度的元组列表转换为数据帧

2 个答案: