在Pandas中，如何检查三个组合的字符串列是否== 10个字符，如果是，则插入新列？

Question

我想将三个Pandas字符串列合并为一个新列，如果新列的总合并字符等于10。

如果不等于10，则检查下一个合并的列。

如果Phone1Area等于3个字符串字符，并且Phone1Prefix等于3个字符串字符，并且Phone1NumberPart等于4个字符串字符，换句话说，总共十个字符，我已经尝试将这些列一起添加到新列中。 我尝试添加列（如果它们是3 + 3 + 4个字符），df.loc和其他更多内容。

这是数据集的示例：

这是代码：

dfp['p1'] = df[(df['Phone1Area'].str.len() == 3.0)]['Phone1Area'] + 
df[(df['Phone1Exchange'].str.len() == 3.0)]['Phone1Exchange'] + 
df[(df['Phone1NumberPart'].str.len() == 4.0)]['Phone1NumberPart']


dfp['p2'] = df[(df['Phone2Area'].str.len() == 3.0)]['Phone2Area'] + 
df[(df['Phone2Exchange'].str.len() == 3.0)]['Phone2Exchange'] + 
df[(df['Phone2NumberPart'].str.len() == 4.0)]['Phone2NumberPart']

df_phone.loc[df_phone['p1'].str.len() == 10, 'phone'] = df_phone['p1']
df_phone.loc[df_phone['p2'].str.len() == 10, 'phone'] = df_phone['p2']

这是我想要的操作，但是在Pandas中：

if df_phone['p1'].str.len() == 10:
    then insert df_phone['p1'] into df_phone['phone']
elif df_phone['p2'].str.len() == 10:
    then insert df_phone['p2'] into df_phone['phone']
elif df_phone['p3'].str.len() == 10:
    then insert df_phone['p3'] into df_phone['phone']

我希望phone栏包含电话1的10个字符，如果不是10个字符，则phone栏包含电话2的10个字符，依此类推。

但是结果之一是：

AttributeError: 'DataFrame' object has no attribute 'str'

任何想法如何解决这个问题？

Answer 1

这应该有助于：

df['phone'] = ''
df['test_phone'] = df['phone1Area'] + df['phone1Exchange'] + df['phone1NumberPart']
df['phone'][df['test_phone'].str.len() == 10] = df['test_phone']
df['test_phone'] = df['phone2Area'] + df['phone2Exchange'] + df['phone2NumberPart']
df['phone'][(df['test_phone'].str.len() == 10) & (df['phone'] == '')] = df['test_phone']
df['test_phone'] = df['phone3Area'] + df['phone3Exchange'] + df['phone3NumberPart']
df['phone'][(df['test_phone'].str.len() == 10) & (df['phone'] == '')] = df['test_phone']
etc.

Answer 2

使用np.select另一种解决方案应该更快：

conditions = [df_phone['p1'].str.len() == 10, df_phone['p2'].str.len() == 10,\
              df_phone['p3'].str.len() == 10]
choices = [df_phone['p1'], df_phone['p2'], df_phone['p3']]

df_phone['phone'] = np.select(conditions, choices, default = '')

文档：

• np.select ：选择条件中遇到的第一个True值的选择。 如果仅为False ，则使用default填充。