【问题标题】:How to map dictionary values to data frame column which has values as lists如何将字典值映射到具有列表值的数据框列
【发布时间】:2020-08-25 18:46:05
【问题描述】:

我有一个数据框:

df = pd.DataFrame(
    {'title':['a1','a2','a3','a4','a5'],
     'genre_name':[
         ['family', 'animation'],
         ['action', 'family', 'comedy'],
         ['family', 'comedy'],
         ['horror','action'],
         ['family', 'animation','comedy']]}
)

df
      title    genre_name
0      a1     ['family', 'animation']
1      a2     ['action', 'family', 'comedy']
2      a3     ['family', 'comedy']
3      a4     ['horror','action]
4      a5     ['family', 'animation','comedy']

我有字典:

dict={'1':'family','2':'animation','3':'action','4':'comedy','5':'horror'}

我想创建一个名为“genre_ids”的新列,它将所有流派名称映射到字典“dict”中的键。

所需的df是:

df
          title    genre_name                       genre_ids
    0      a1     ['family', 'animation']           [1,2]
    1      a2     ['action', 'family', 'comedy']    [3,1,4]
    2      a3     ['family', 'comedy']              [1,4]
    3      a4     ['horror','action]                [5,3]
    4      a5     ['family', 'animation','comedy']  [1,2,4]

我怎样才能做到这一点?

【问题讨论】:

    标签: python pandas dataframe dictionary


    【解决方案1】:

    将字典名称从dict 更改为另一个变量,因为内置(python 代码字),然后在列表理解中将键与值交换并映射值:

    d={'1':'family','2':'animation','3':'action','4':'comedy','5':'horror'}
    
    d1 = {v:k for k, v in d.items()}
    df['genre_ids'] = df['genre_name'].apply(lambda x: [d1.get(y) for y in x])
    #alternative
    #df['genre_ids'] = [[d1.get(y) for y in x] for x in df['genre_name']]
    print (df)
      title                   genre_name  genre_ids
    0    a1          [family, animation]     [1, 2]
    1    a2     [action, family, comedy]  [3, 1, 4]
    2    a3             [family, comedy]     [1, 4]
    3    a4             [horror, action]     [5, 3]
    4    a5  [family, animation, comedy]  [1, 2, 4]
    

    编辑:您还可以指定如果不匹配会发生什么,这里为第一个列表添加crime:

    df = pd.DataFrame({'title':['a1','a2','a3','a4','a5'], 
                       'genre_name':[['crime', 'animation'],['action', 'family', 'comedy'],
                                     ['family', 'comedy'],['horror','action'],
                                     ['family', 'animation','comedy']]})
    
    d={'1':'family','2':'animation','3':'action','4':'comedy','5':'horror'}
    
    
    d1 = {v:k for k, v in d.items()}
    #no matched values repalced to None
    df['genre_ids0'] = df['genre_name'].apply(lambda x: [d1.get(y) for y in x])
    #no match replaced to default value
    df['genre_ids1'] = df['genre_name'].apply(lambda x: [d1.get(y, 0) for y in x])
    #no match is removed
    df['genre_ids2'] = df['genre_name'].apply(lambda x: [d1[y] for y in x if y in d1])
    print (df)
      title                   genre_name genre_ids0 genre_ids1 genre_ids2
    0    a1           [crime, animation]  [None, 2]     [0, 2]        [2]
    1    a2     [action, family, comedy]  [3, 1, 4]  [3, 1, 4]  [3, 1, 4]
    2    a3             [family, comedy]     [1, 4]     [1, 4]     [1, 4]
    3    a4             [horror, action]     [5, 3]     [5, 3]     [5, 3]
    4    a5  [family, animation, comedy]  [1, 2, 4]  [1, 2, 4]  [1, 2, 4]
    

    【讨论】:

    • 您创建d1 的目的是什么?
    • @BearBrown - 交换键 -> 值
    猜你喜欢
    • 1970-01-01
    • 1970-01-01
    • 2018-09-21
    • 1970-01-01
    • 2015-07-24
    • 2021-05-26
    • 2021-10-26
    • 2021-07-14
    相关资源
    最近更新 更多