在 python 中使用 apply 或 iterrows 而不是 groupby？

Question

df = pd.DataFrame([[15000, 2015], [20000,2015], [25000,2015], [15000, 2016], [20000,2016], [25000,2016], [10500, 2017], [54000,2017], [34000,2017]], columns=['income', 'year'])


income          year
15000           2015
20000           2015
25000           2015
19000           2016
36000           2016
20000           2016
10500           2017
54000           2017
34000           2017

Hello,

If I have a dataframe like the one above and I want to loop through each year in python and create a median income value for each year, how would I go about it?

Would the apply function or the groupby function be best?

I can get this to work:

df.groupby(df.year)[['income']].median()

I was wondering whether there was an alternative such as apply or iterrows?

Many thanks.

Answer 1

df.groupby is the best way to go when you are doing certain aggregations.

This is the right way to use it:

In [85]: df.groupby('year', as_index=False)['income'].median()
Out[85]: 
   year  income
0  2015   20000
1  2016   20000
2  2017   34000

After OP's comment:

In [239]: res = df.groupby('year', as_index=False)['income'].median()
In [259]: d = res.set_index('year').to_dict()['income']

Then you can query the above dict d to get mean for a certain year, like this:

In [268]: d.get(2015)
Out[268]: 20000

In [269]: d.get(2016) 
Out[269]: 20000

In [270]: d.get(2017) 
Out[270]: 34000