计算熊猫数据框中行之间的差异

Question

I'd like to compute the difference between two categories in a dataframe. For example, in the following case, I want to compute the differences between male and female on each job. However, there are some jobs done by only male or female. What is an efficient way to do it? Thanks.

import pandas as pd

pd.DataFrame({'job': ['a', 'a', 'b', 'b', 'c'], 'gender':['M', 'F', 'M', 'F', 'M'], 'income':[300, 200, 450, 400, 350]})

Out[3]: 
  gender  income job
0      M     300   a
1      F     200   a
2      M     450   b
3      F     400   b
4      M     350   c

Answer 1

You could do a pivot such that the male and female pay for the same job are on the same row. Then you can visually compare, or run other row-based code.

import pandas as pd

df = pd.DataFrame({'job': ['a', 'a', 'b', 'b', 'c'], 'gender':['M', 'F', 'M', 'F', 'M'], 'income':[300, 200, 450, 400, 350]})

compare_income_by_gender_df = df.pivot(index='job', columns='gender', values='income')

print compare_income_by_gender_df

resulting in

python pivot.py
gender    F    M
job             
a       200  300
b       400  450
c       NaN  350

see also: Pandas Reshaping and Pivot Tables