用于提取每组最近 2 行的数据框

Question

A simple data-frame and I want to pick the most recent 2 rows (sorted by "Year") with all columns.

import pandas as pd

data = {'People' : ["John","John","John","Kate","Kate","David","David","David","David"],
'Year': ["2018","2019","2006","2017","2012","2006","2019","2018","2017"],
'Sales' : [120,100,60,150,135,140,90,110,160]}

df = pd.DataFrame(data)

I tried below but it doesn't produce what's wanted:

df = df.groupby('People')
df_1 = pd.concat([df.head(2)]).drop_duplicates().sort_values('Year').reset_index(drop=True)

What's the right way to write it? Thank you.

Answer 1

IIUC, use pandas.DataFrame.nlargest :

df['Year'] = df['Year'].astype(int)
df.groupby('People', as_index=False).apply(lambda x: x.nlargest(2, "Year"))

Output:

    People  Year  Sales
0 6  David  2019     90
  7  David  2018    110
1 1   John  2019    100
  0   John  2018    120
2 3   Kate  2017    150
  4   Kate  2012    135