Question

我有一个 Given_Col ，其中有5个不同的值。我需要创建新列，如下所示：

A和B都去了---> col1

A和B都不走---> col1

A走B不走---> col2

A不会走，B会走---> col2

不知道--->如果值均为NaN两者都

      Given_Col                    Expexted_Col1              Expexted_Col2    

Both A and B goes                 Both A and B goes               No idea
Neither A nor B goes             Neither A nor B goes             No idea
A goes B doesn't go                    No idea               A goes B doesn't go  
A doesn't go B goes                    No idea               A doesn't go B goes 
A goes B doesn't go                    No idea               A goes B doesn't go
Neither A nor B goes             Neither A nor B goes             No idea 
No idea                                No idea                    No idea 
Both A and B goes                  Both A and B goes              No idea

我想不出任何解决方案。实际的方法是什么？

注意：我考虑过复制现有列并映射值吗？

Answer 1

我认为应该有两个条件列分配。

每个人都根据选择标准为该列选择有效的条目。如果您有五种以上的可能性，这可能会很笨拙，但是对于这种情况，它应该可以很好地工作。

df['Expexted_Col1'] = df.apply(lambda x: x['Given_Col'] if (x['Given_Col'] == 'Both A and B goes' or x['Given_Col'] == 'Neither A nor B goes') else 'No idea', axis = 1)
df['Expexted_Col2'] = df.apply(lambda x: x['Given_Col'] if (x['Given_Col'] == "A goes B doesn't go" or x['Given_Col'] == "A doesn't go B goes") else 'No idea', axis = 1)

Answer 2

将pandas.DataFrame.assign与fillna结合使用的一种方式：

mapper = {'col1': ['Both A and B goes', 'Neither A nor B goes'],
 'col2': ["A goes B doesn't go", "A doesn't go B goes"]}

s = df["Given_Col"]
new_df = df.assign(**{k: s[s.isin(v)] for k, v in mapper.items()}).fillna("No idea")
print(new_df)

输出：

              Given_Col                  col1                 col2
0     Both A and B goes     Both A and B goes              No idea
1  Neither A nor B goes  Neither A nor B goes              No idea
2   A goes B doesn't go               No idea  A goes B doesn't go
3   A doesn't go B goes               No idea  A doesn't go B goes
4   A goes B doesn't go               No idea  A goes B doesn't go
5  Neither A nor B goes  Neither A nor B goes              No idea
6               No idea               No idea              No idea
7     Both A and B goes     Both A and B goes              No idea

Answer 3

您可以使用几个np.where函数来做到这一点：

df['col1'] = np.where(df['Given_Col'] == 'Both A and B goes', 'Both A and B goes', df['col1'])
df['col2'] = np.where(df['Given_Col'] == 'Both A and B goes', 'No idea', df['col1'])
df['col1'] = np.where(df['Given_Col'] == 'Neither A nor B goes', 'Neither A nor B goes', df['col2'])
df['col2'] = np.where(df['Given_Col'] == 'Neither A nor B goes', 'No idea', df['col2'])

您可以从这里继续。...

按值将一列以大熊猫的形式分为两列

3 个答案: