使用字典在pandas列中填充NaN

Question

Is there a way to map column values using a dictionary that does not include all columns. 有没有办法使用不包含所有列的字典映射列值。 Eg: 例如：

Let's say my dataframe is: 假设我的数据框是：

 A    B    C    D    E    F
nan  nan  nan  nan  nan  nan

and I have a dictionary which I would like to use as a mapper: 我有一本字典，我想用作映射器：

d = {'A': 1, 'B': 1, 'E': 1}

so the output should be replacing by 0 those values that are not in the dictionary 所以输出应该用0替换那些不在字典中的值

A    B    C    D    E    F
1    1    0    0    1    0

Answer 1

The most idiomatic choice is with two fillna calls, 最惯用的选择是两个fillna调用，

df.fillna(d).fillna(0, downcast='infer')
df

   A  B  C  D  E  F
0  1  1  0  0  1  0

piRSquared suggests assign as an alternative to the first fillna call, piRSquared建议assign作为第一个fillna调用的替代方案，

df.assign(**d).fillna(0, downcast='infer')
df

   A  B  C  D  E  F
0  1  1  0  0  1  0

Another option is to use Index.isin on the columns. 另一种选择是在列上使用Index.isin 。 This is the single row form: 这是单行形式：

df[:] = [df.columns.isin(d.keys()).astype(int)]

To generalise to N rows, we use repeat : 要概括为N行，我们使用repeat ：

df[:] = df.columns.isin(d.keys()).astype(int)[None,:].repeat(len(df), axis=0)
df

   A  B  C  D  E  F
0  1  1  0  0  1  0

For fun, you can also use reindex : 为了好玩，您还可以使用reindex ：

pd.DataFrame(d, index=df.index).reindex(df.columns, axis=1, fill_value=0)

   A  B  C  D  E  F
0  1  1  0  0  1  0