如何在基于另一列的列中填充缺失值

Question

I have a dataframe called shoes

Brand   Comment
Ugg       NaN
Prada     NaN
Clarks    NaN
Ugg       NaN
Clark     NaN
Prada     Made from horse leather
Prada     Made from pig leather
Prada     NaN
Ugg       Made from Australian cow leather
...

and another dataframe df_mode which was obtained by taking the mode of the comments for each shoe brand in the shoes dataframe for nonnull values

Brand  Comment
Ugg    Made from sheep 
Prada  Made from pig leather
Clarks Made from Cow leather

How can I assign the missing values for each shoe brand in the shoes dataframe with its respective mode response shown in the df_mode dataframe.

This is basically what I'm trying to achieve

Brand   Comment
Ugg       Made from sheep
Prada     Made from pig leather
Clarks    Made from Cow leather
Ugg       Made from sheep
Clark     Made from Cow leather
Prada     Made from horse leather
Prada     Made from pig leather
Prada     Made from pig leather
Ugg       Made from Australian cow leather

Answer 1

使用np.where

shoes['Comment']=np.where(shoes['Comment'].isnull(),shoes['Brand'].map(dict(zip(df_mode['Brand']))),df_mode['Comment'],shoes['Comment'])

Answer 2

使用loc和map ：

shoes.loc[shoes.Comment.isna(), 'Comment'] = shoes.Brand.map(df_mode.set_index('Brand')['Comment'])

Answer 3

you can first groupby by Brand column, then fill the missing values. here is the implementation:

df['Comment'] = df.groupby(['Brand'], sort=False)['Comment'].apply(lambda x: x.ffill().bfill())