使用字典替换列值

时间:2017-08-31 05:44:21

标签: python pandas dictionary dataframe replace

我有这个数据框,其中性别应该是男性或女性。

from io import StringIO
import pandas as pd

audit_trail = StringIO('''
course_id AcademicYear_to months TotalFee Gender
260 2017 24 100 male
260 2018 12 140 male
274 2016 36 300 mail
274 2017 24 340 female
274 2018 12 200 animal
285 2017 24 300 bird
285 2018 12 200 maela
''')

df11 = pd.read_csv(audit_trail, sep=" "  )

我可以使用字典来纠正拼写错误。

corrections={'mail':'male', 'mael':'male', 'maae':'male'}
df11.Gender.replace(corrections)

但是我正在寻找一种方法来保持只有男性/女性和其他"其余选项的类别。预期产出:

0      male
1      male
2      male
3    female
4    other
5    other
6      male
Name: Gender, dtype: object

2 个答案:

答案 0 :(得分:3)

corrections字典中添加另外两个虚拟条目:

corrections = {'male'   : 'male',    # dummy entry for male
               'female' : 'female',  # dummy entry for female
               'mail'   : 'male', 
               'maela'  : 'male', 
               'maae'   : 'male'}

现在,使用mapfillna

df11.Gender = df11.Gender.map(corrections).fillna('other')
df11

   course_id  AcademicYear_to  months  TotalFee  Gender
0        260             2017      24       100    male
1        260             2018      12       140    male
2        274             2016      36       300    male
3        274             2017      24       340  female
4        274             2018      12       200   other
5        285             2017      24       300   other
6        285             2018      12       200    male

答案 1 :(得分:2)

您可以使用:

corrections={'mail':'male', 'maela':'male', 'maae':'male', 'male':'male', 'female':'female'}
df11[['Gender']] = df11[['Gender']].applymap(corrections.get).fillna('other')
print (df11)
   course_id  AcademicYear_to  months  TotalFee  Gender
0        260             2017      24       100    male
1        260             2018      12       140    male
2        274             2016      36       300    male
3        274             2017      24       340  female
4        274             2018      12       200   other
5        285             2017      24       300   other
6        285             2018      12       200    male

编辑:

对于替换,只有一列更好cᴏʟᴅsᴘᴇᴇᴅ的答案。如果要替换多个列,最好是applymap