你可以像这样实现它

import pandas as pd
# Example data
data = {
    'Year': [2019, 2019, 2022, 2022, 2021, 2021],
    'Tm': ['A', 'B', 'C', 'D', 'E', 'F'],
    'Lg': ['NORW', 'ABC', 'NORW', 'ABC', 'LAW', 'ABC'],
    'Lev': ['A-', 'Rk', 'A+', 'A-', 'A', 'A+'],
    'PA': [117, 100, 358, 200, 9, 50]
}
prospect = pd.DataFrame(data)
levels = ['Rk', 'A-', 'A', 'A+', 'AA', 'AAA', 'MLB']
# Convert 'Lev' to categorical type with the specified order
prospect['Lev'] = pd.Categorical(prospect['Lev'], categories=levels, ordered=True)
# Filter and get the first row for each year after sorting in descending order by 'PA'
prospect = prospect.sort_values(['Year', 'PA'], ascending=[True, False]).groupby('Year').first().reset_index()
print(prospect)

字符串
输出量：

Year Tm    Lg Lev   PA
0  2019  A  NORW  A-  117
1  2021  F   ABC  A+   50
2  2022  C  NORW  A+  358

型

你的例子

import pandas as pd
prospect = pd.read_html('https://www.baseball-reference.com/register/player.fcgi?id=bishop000hun')[0]
levels = ['Rk', 'A-', 'A', 'A+', 'AA', 'AAA', 'MLB']
prospect = prospect[['Year', 'Tm', 'Lg', 'Lev', 'PA']][prospect['Lev'].isin(levels)]

字符串
使用键的sort_values

m = {j: i for i, j in enumerate(levels)} 
out = prospect.sort_values('Lev', key=lambda x: x.map(m)).groupby(['Year']).tail(1)

型
输出：

Year    Tm              Lg      Lev PA
6   2019    Salem-Keizer    NORW    A-  117
10  2021    Eugene          HAW     A+  15
15  2022    Eugene          NORW    A+  358

型