pandas：用列表替换列中的非空行答案

【问题标题】：pandas: replace non-empty rows in a column with a listpandas：用列表替换列中的非空行
【发布时间】：2020-10-14 18:16:05
【问题描述】：

我有一个如下数据框（但更长），其中一些行包含 None 值：

data = {'first Column':  ['she is', 'they are',NaN,'we are',NaN],
    'second Column ': ['my', 'her',NaN,'his',NaN],
    'third column': ['friend', 'brothers',NaN,'sisters',NaN]
    }

df = pd.DataFrame (data, columns = ['first Column','second Column','third column])

my_list= ['gold','silver','bronze']

如果“第三列”中的每一行不包含任何值，我想用一个列表替换它，如下所示：

  desired output:
     first column   second column   third column
  0  she is          my             ['gold','silver','bronze']
  1  they are        her            ['gold','silver','bronze']
  2  NaN             NaN            NaN
  3  we are          his            ['gold','silver','bronze']
  4  NaN             NaN            NaN

我尝试过 np.where，但它没有选择所需的行

   np.where(df.loc[df['third column'] != 'NaN', [','.join(my_list)], df['third column']

【问题讨论】：

标签： python pandas list replace

【解决方案1】：

df.loc[df['third column'].notna(),'third column']=[my_list]

结果：

  first Column second Column             third column
0       she is             my  [gold, silver, bronze]
1     they are            her  [gold, silver, bronze]
2          NaN            NaN                     NaN
3       we are            his  [gold, silver, bronze]
4          NaN            NaN                     NaN

PS：如果您将 data 指定为带有列名的 dict，则不需要在数据框构造函数中使用 columns 参数。

【讨论】：