Question

我有一列带有客户名称的列，当客户有2种产品时，该列将重复。我必须创建一个新状态，以根据情况将客户状态分组为一个。因此，我必须将客户X与另一个X进行比较，以产生一个新的状态。

Customer|Status  |Cancaled_at|new status
X       |Active  |-          |
X       |Canceled|2019-xx-xx |
Y       |Active  |-          |
Z       |Active  |-          |
A       |Canceled|-          |

所需的输出：

Customer|Status  |Cancaled_at|new status
X       |Active  |-          |Canceled
X       |Canceled|2019-xx-xx |Canceled
Y       |Active  |-          |
Z       |Active  |-          |
A       |Canceled|-          |

Answer 1

有一种简单的方法可以找到熊猫中所有重复的值：

df['new_status'][(df.duplicated('Customer', False))] = 'Canceled'

这使new_status列Canceled的位置数据框的“客户”列具有重复的值。

Answer 2

我认为您需要：

df = pd.DataFrame({'Customer':['X','X','Y','Z','A'], 'status':['active','canceled','active','active','canceled'],
    'Canceled_at':[None, '2019-01-01', None, None,None]})


df['new_status'] = np.where((df['status']=='canceled') & (~df['Canceled_at'].isnull()), 'canceled', None)
df['new_status'] = df.groupby('Customer')['new_status'].bfill()

print(df)

输出：

 Canceled_at   Customer    status new_status                                                                                           
0        None        X    active   canceled                                                                                           
1  2019-01-01        X  canceled   canceled                                                                                           
2        None        Y    active       None                                                                                           
3        None        Z    active       None                                                                                           
4        None        A  canceled       None

Answer 3

此代码使用sort_values()，fillna()和shift()：

df = df.sort_values(by=['Customer', 'Status'])
df['new_status'] = df[df.Status == 'Canceled']['Status']
df.loc[((df['Customer'] != df['Customer'].shift(-1)) & (df['new_status'].isnull())), 'new_status'] = ''
df['new_status'].fillna(method = 'backfill', limit = 1, inplace = True)
df.sort_index(inplace = True)

产生以下输出：

    Customer    Status      Cancaled_at new_status
0   X           Active      -           Canceled
1   X           Canceled    2019-xx-xx  Canceled
2   Y           Active      -   
3   Z           Active      -   
4   A           Canceled    -           Canceled

比较熊猫列中的两个重复值

3 个答案: