finding the minimum value of matched rows between two dataframes

I have Two data frames

import pandas as pd
exam_1 = pd.DataFrame({'user': ['A', 'B', 'C'], 
                   'marks': [10, 50, 40]})

exam_2 = pd.DataFrame({'user': ['A', 'C', 'D'], 
                   'marks': [30, 20, 30]})

I’m trying to get a resultant data frame, with users appeared in both exams with lowest marks.
In the above example result should be:

          user   marks
           A      10
           C      30

I did write following lines to get users who appeared in both exams but, stuck in getting the lowest marks. Here is what I have tried

tmp = pd.merge(exam_1, exam_2, how='inner', on=['user'])
USERS = exam_1[ exam_1['user'].isin(tmp['user'])]

can someone please help me to get lowest exam marks in the resultant dataframe?

>Solution :

One option could be to set "user" as index, then use numpy.minimum and dropna:

import numpy as np

out = np.minimum(exam_1.set_index('user'),
                 exam_2.set_index('user')).dropna()

Output:

      marks
user       
A      10.0
C      20.0

With your merge approach:

out = (pd.merge(exam_1, exam_2, on='user')
         .set_index('user')
         .filter(like='marks')
         .min(axis=1)
      )

Output:

user
A    10
C    20
dtype: int64

Another, more verbose, option with concat:

common = set(exam_1['user']) & set(exam_2['user'])
m1 = exam_1['user'].isin(common)
m2 = exam_2['user'].isin(common)

out = (pd.concat([exam_1[m1], exam_2[m2]])
         .groupby('user', as_index=False).min()
      )

Output:

  user  marks
0    A     10
1    C     20