pandas 如何在熊猫中定义用户定义的函数
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/35414431/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
How to define user defined function in pandas
提问by Edwin Baby
I have a csv file that contains information like
我有一个包含以下信息的 csv 文件
name salary department
a 2500 x
b 5000 y
c 10000 y
d 20000 x
I need to convert this using Pandas to the form like
我需要使用 Pandas 将其转换为类似的形式
dept name position
x a Normal Employee
x b Normal Employee
y c Experienced Employee
y d Experienced Employee
if the salary <=8000 Position is Normal Employee
如果薪水 <=8000 职位是普通员工
if the salary >8000 && <=25000 Position is Experienced Employee
如果薪水 >8000 && <=25000 职位是有经验的员工
My default code for group by
我的默认分组代码
import csv
import pandas
pandas.set_option('display.max_rows', 999)
data_df = pandas.read_csv('employeedetails.csv')
#print(data_df.columns)
t = data_df.groupby(['dept'])
print t
What are the changes i need to make in this code to get the output that i mentioned above
我需要在此代码中进行哪些更改才能获得我上面提到的输出
采纳答案by EdChum
You could define 2 masks and pass these to np.where:
您可以定义 2 个掩码并将它们传递给np.where：
In [91]:
normal = df['salary'] <= 8000
experienced = (df['salary'] > 8000) & (df['salary'] <= 25000)
df['position'] = np.where(normal, 'normal emplyee', np.where(experienced, 'experienced employee', 'unknown'))
df
Out[91]:
name salary department position
0 a 2500 x normal emplyee
1 b 5000 y normal emplyee
2 c 10000 y experienced employee
3 d 20000 x experienced employee
Or slightly more readable is to pass them to loc:
或者稍微更具可读性的是将它们传递给loc：
In [92]:
df.loc[normal, 'position'] = 'normal employee'
df.loc[experienced,'position'] = 'experienced employee'
df
Out[92]:
name salary department position
0 a 2500 x normal employee
1 b 5000 y normal employee
2 c 10000 y experienced employee
3 d 20000 x experienced employee
回答by Fabio Lamanna
I would use a simple function like:
我会使用一个简单的函数，如：
def f(x):
if x <= 8000:
x = 'Normal Employee'
elif 8000 < x <= 25000:
x = 'Experienced Employee'
return x
and then apply it to the df:
然后将其应用于df：
df['position'] = df['salary'].apply(f)
回答by IanS
A useful function is apply:
一个有用的功能是apply：
data_df['position'] = data_df['salary'].apply(lambda salary: 'Normal Employee' if salary <= 8000 else 'Experienced Employee', axis=1)
This applies the lambdafunction to every element in the salary column.
这将lambda函数应用于工资列中的每个元素。
