使用numpy.where或numpy.select：

df['specialty'] = np.select([df['specialty'].str.contains(','),
                             df['specialty'].eq('missing')], [7, 6], 
                             default=df['specialty'])

df['specialty'] = np.where(df['specialty'].str.contains(','), 
                           7, 
                           df['specialty'].replace({'missing':6}))

print (df)
  specialty
0         1
1         7
2         2
3         6
4         6
5         1
6         3
7         7
8         4
9         1

对loc使用布尔索引：

df.loc[df['specialty'].str.contains(','), 'specialty'] = 7
df.loc[df['specialty'].eq('missing'), 'specialty'] = 6

输出：

specialty
0         1
1         7
2         2
3         6
4         6
5         1
6         3
7         7
8         4
9         1

您可能还想从真实的值（2+5 -〉7）计算：

df['specialty'] = [6 if s=='missing' else sum(map(int, s.split(',')))
                   for s in df['specialty']]

或者：

df['specialty'] = (df['specialty'].replace({'missing': '6'})
                   .str.split(',', expand=True).astype(float)
                   .sum(axis=1).astype(int)
                  )

您可以用途：

df['spe2'] = (df['specialty'].astype(str).replace({'missing': '6'})
                             .str.split(',').explode().astype(int)
                             .groupby(level=0).sum())
print(df)

# Output
  specialty  spe2
0         1     1
1       2,5     7
2         2     2
3         6     6
4   missing     6
5         1     1
6         3     3
7     1,3,4     8
8         4     4
9         1     1