如何从数据框创建嵌套字典

How to create a nested dictionary from a dataframe

我有一个这样的数据框:

df = pd.DataFrame(id:{1,2,1,4,4},
course:{math,math,sci,art,math},
result:{pass,pass,fail,fail,fail}}

我想像这样创建一个嵌套字典: 对于每个 ID,我想制作一个包含已通过课程和未通过课程的嵌套字典。

{id:{pass:{courses},fail:{courses}}}

{1:{pass:{math},fail:{sci}},2:{pass:{math}},4:{fail:{art,math}}}

非常感谢

试试这个:

df = pd.DataFrame({'id':[1,2,1,4,4],
'course':['math','math','sci','art','math'],
'result':['pass','pass','fail','fail','fail']})


df.groupby(['result', 'id'])['course'].agg(list).unstack().to_dict()

输出:

{1: {'fail': ['sci'], 'pass': ['math']},
 2: {'fail': nan, 'pass': ['math']},
 4: {'fail': ['art', 'math'], 'pass': nan}}

嗯,是的,在@mozway 解决方案中达到顶峰,使用 set 而不是 list:

df.groupby(['result', 'id'])['course'].agg(set).unstack().to_dict()

输出:

{1: {'fail': {'sci'}, 'pass': {'math'}},
 2: {'fail': nan, 'pass': {'math'}},
 4: {'fail': {'art', 'math'}, 'pass': nan}}

假设输入:

   id course result
0   1   math   pass
1   2   math   pass
2   1    sci   fail
3   4    art   fail
4   4   math   fail

你可以使用嵌套的groupby:

out = (df.groupby('id')
         .apply(lambda g: g.groupby('result')['course']
                           .agg(set).to_dict())
         .to_dict()
       )

输出:

{1: {'fail': {'sci'}, 'pass': {'math'}},
 2: {'pass': {'math'}},
 4: {'fail': {'art', 'math'}}}

或枢轴table:

(df.pivot_table(columns='id', index='result', values='course', aggfunc=set)
   .to_dict()
)

输出:

{1: {'fail': {'sci'}, 'pass': {'math'}},
 2: {'fail': nan, 'pass': {'math'}},
 4: {'fail': {'art', 'math'}, 'pass': nan}}