Pandas df.mode,列中每个单元格有多个值

首次使用 Series.str.split 具有 expand=True 用于数据帧和重塑 DataFrame.stack ,然后按 Series.value_counts 并通过 Series.head :

df = pd.DataFrame({'Keywords':['aa,bb,vv,vv','aa,aa,cc,bb','zz,bb,aa,ss']})

N = 5
df1 = (df.Keywords.str.split(',', expand=True)
         .stack()
         .value_counts()
         .head(N)
         .rename_axis('val')
         .reset_index(name='count'))
print (df1)
  val  count
0  aa      4
1  bb      3
2  vv      2
3  zz      1
4  cc      1

如果没有丢失的值,另一个解决方案是展平拆分的列表并按 Counter :

from collections import Counter
N = 5
df1 = pd.DataFrame(Counter([y for x in df.Keywords for y in x.split(',')]).most_common(N), 
                   columns=['val','count'])
print (df1)
  val  count
0  aa      4
1  bb      3
2  vv      2
3  zz      1
4  cc      1