如何根据条件拆分列答案

【问题标题】：how to split column with condition如何根据条件拆分列
【发布时间】：2018-12-09 03:37:21
【问题描述】：

所以，我想知道如何根据条件划分列。我的想法是研究用户的活动，但为此我需要设置一个条件。我有数据框：

df = pd.DataFrame({'User': ["juan","juan","juan","juan","petter","petter","petter","petter","petter","petter","petter","petter","ana","ana","ana","ana","raul","raul","raul","raul"],
               'time': ["2/1/2019","3/1/2019","4/1/2019","6/1/2019","2/1/2019","5/1/2019","6/1/2019","10/1/2019","11/1/2019","12/1/2019","13/1/2019","14/1/2019","8/1/2019","10/1/2019","15/1/2019","20/1/2019","15/1/2019","17/1/2019","18/1/2019","19/1/2019"],
                'activity': ["fly", "hotel","car","jump","fly", "hotel","jump","car","fly", "car","hotel","car","car", "hotl","car","hotel","fly", "hotel","car","car"],
              '%timeper_user': ["4 days","4 days","4 days","4 days","8 days","8 days","8 days","8 days","3 days","3 days","3 days","3 days","12 days","12 days","12 days","12 days","4 days","4 days","4 days","4 days"]})

如您所见，每个用户都有一个列（时间），每个用户还有一个列（%timeper_user）。然后是一个列（活动），它是每个用户在一段时间内执行的活动。这个想法是对不同列中的每个活动进行“条件拆分”。第一幕，第二幕，第三幕，第三幕。但是当用户在时间之外（time +% timeper_user）执行活动时，将活动放在不同的列中，例如：Act21，Act 22，Act 23，Act24 我希望它是这样的：

df2 = pd.DataFrame({'User': ["juan","petter","ana","raul"],
              "act1":["fly","fly","car","fly"],
              "act2":["hotel","hotel","hotel","hotel"],
              "act3":["car","jump","car","car"],
              "act4":["jump","car","hotel","car"],
              "actn":["","","",""],
              "act21":["","fly","",""],
              "act22":["","car","",""],
              "act23":["","hotel","",""],
              "act24":["","car","",""]})

(DF2) 是我想要的输出见用户 Petter 超出时间（2/1/2019 + 8 天）= 10/1/2019。因此，从 2019 年 11 月 1 日起，活动将放在 Act21、Act22、Act23、Act24 中。我有很多用户，所以我不知道如何执行一个执行此操作并获取所有内容的功能（逐个用户）。如果你能帮助我，我将不胜感激。谢谢

【问题讨论】：

欢迎来到 SO，您能否在您的帖子中也发布示例预期输出。
@RavinderSingh13 抱歉，如果我解释得不好。但是我发布了它，DF是我拥有的数据。第二种情况（DF2）是我想要的输出
是否正确，如果满足这两个条件，该活动应出现在 Act21、Act22 等中：1) 给定用户在过去执行此活动 2) time 重复活动更大之前活动的time + 之前活动的timeper_user？
@AlexeyMilogradov 是的

标签： python database python-3.x pandas split

【解决方案1】：

这个想法是。如果用户在范围（每个用户的时间+％时间）之间进行事件，则意味着所有活动都属于进入活动范围 1（act1、Act12、Act13、Act 14）。如果日期较大，则表示用户将输入活动 2（act21、act22、act23、act24）。简单来说......如果petter从美国去马德里，他可能会去酒店，租一辆车，然后尝试飞行。但是当他回到美国时，Petter 可能会在那里购买第二次航班（这将进入范围活动 2（第 21 幕，第 22 幕，第 23 幕，第 24 幕）。如果你运行 df... 是我拥有的数据框。 . 和 df2 是我想要制作的数据

【讨论】：