通过不同列的重复值之间的条件

Question

当一个客户有多个订阅时，它就是重复的。我想为整个客户状态生成一个 new_status，而不是为每个订阅生成一个 new_status：给已重新激活订阅的客户以及已取消一项订阅但仍有另一项有效订阅的客户。

df:

Customer | Status  | Canceled_at | Created  | New_status
 X       | Active  |             |8/9/2017  |
 X       |Canceled |  8/3/2017   |6/19/2017 |             
 Y       | Active  |             |2/13/2019 |
 Y       |Canceled | 11/28/2018  |10/14/2018|
 Z       | Active  |             |3/29/2018 |
 Z       |Canceled | 8/8/2018    |7/10/2018 |
 A       |Canceled | 9/2/2018    |7/10/2018 |          
 A       |Canceled | 9/29/2018   |7/12/2018 |
 A       |Active   |             |5/31/2018 |

这些情况的条件是：如果取消副本的 'canceled_at' 日期 > 活动的 'created' 日期：新_status 将是 'Downgrade' 如果取消副本的 'canceled_at' 日期 < 'created' 日期活动：new_status 将是 'Reactivate'

期望的输出：

Customer | Status  | Canceled_at | Created  | New_status
 X       | Active  |             |8/9/2017  |Reactivate
 X       |Canceled |  8/3/2017   |6/19/2017 |Reactivate              
 Y       | Active  |             |2/13/2019 |Reactivate
 Y       |Canceled | 11/28/2018  |10/14/2018|Reactivate
 Z       | Active  |             |3/29/2018 |Downgrade
 Z       |Canceled | 8/8/2018    |7/10/2018 |Downgrade
 A       |Canceled | 9/2/2018    |7/10/2018 |Downgrade           
 A       |Canceled | 9/29/2018   |7/12/2018 |Downgrade
 A       |Active   |             |5/31/2018 |Downgrade

Answer 1

我太新了，无法发表评论，但我需要更多信息，为什么 'Y' 客户重新激活？也许我不明白你的解释，因为客户 'A' 处于类似情况，而你给了它 'Downgrade'。也许只是 re-type 你的问题，但假装它是给一个 8 岁的孩子阅读的（我）。

这是您想要的代码，它有效：

#convert columns to dates
df['Canceled_at'] = pd.to_datetime(df['Canceled_at'])
df['Created'] = pd.to_datetime(df['Created'])

#make customer a list so we can loop through it
customer = list(df['Customer'].drop_duplicates())

#super awesome for loop that give us the largest date (this is the part where maybe your logic is different than what I read it as)
for c in customer:
    df.loc[(df['Customer'] == c), 'Most Recent Cancel'] = df.loc[(df['Customer'] == c)]['Canceled_at'].max()
    df.loc[(df['Customer'] == c), 'Most Recent Created'] = df.loc[(df['Customer'] == c)]['Created'].max()

#Make 'New_status' column
df.loc[(df['Most Recent Created'] > df['Most Recent Cancel']), 'New_status'] = 'Reactivate'
df.loc[(df['New_status'] != 'Reactivate'), 'New_status'] = 'Downgrade'

通过不同列的重复值之间的条件

Conditional between duplicated values through different columns

python

dataframe

pandas

pandas-loc