如何获取pandas中两列之间的日期范围信息

Question

我有一列日期，我将其分成两列日期， DATE_1和DATE_2 。 我一直试图找到一种方法来获取每个日期的年份周数和该范围内每个日期的星期几，不包括DATE_2例如：

Date_1      Date_2
2020-09-27  2020-10-01
2020-12-24  2020-12-29
2020-12-24  2021-01-03
2020-12-28  2021-01-03

我想得到

Date_1      Date_2      Week                       Days
2020-09-27  2020-10-01  39,40,40,40                Sun,Mon,Tues,Wed
2020-12-24  2020-12-29  52,52,52,53                Thurs,Fri,Sat,Sun,Mon
2020-12-24  2021-01-03  52,52,52,53,53,53,53,53,53 Thurs,Fri,Sat,Sun,Mon,Tues,Wed,Thur,Fri,Sat
2020-12-28  2021-01-03  53,53,53,53,53,53          Mon,Tues,Wed,Thur,Fri,Sat

天数的显示方式可以是全名也可以是对应特定天数的数值，最重要的是数据存在于某处。

我知道熊猫有 date_range 但我不知道如何将它合并到我正在寻找的内容中。 也许它不是熊猫特有的，我不太确定。 任何帮助，将不胜感激。

Answer 1

在评论中的链接的帮助下，我想出了一个使用date_range的解决方案：

import pandas as pd

x = {'Date_1': {0: '2020-09-27', 1: '2020-12-24', 2: '2020-12-24', 3: '2020-12-28'},
    'Date_2': {0: '2020-10-01', 1: '2020-12-29', 2: '2021-01-03', 3: '2021-01-03'}}

weekdays = {1: "Mon", 2: "Tues", 3: "Wed", 4: "Thur", 5: "Fri", 6: "Sat", 7: "Sun"}

df = pd.DataFrame(x)

# Creates a new column containing all the days between Date_1 and Date_2
df["Week"] = df.apply(lambda row: pd.date_range(start=row["Date_1"], end=row["Date_2"], freq="D"), axis=1)
# Using the days, we collect the weekdays of the days
df["Days"] = df["Week"].apply(lambda dates: [weekdays.get(date.isocalendar()[2]) for date in dates])
# Finally we gather the week-number for all of the days
df["Week"] = df["Week"].apply(lambda dates: [date.isocalendar()[1] for date in dates])

输出：

       Date_1      Date_2                                          Week                                                        Days
0  2020-09-27  2020-10-01                          [39, 40, 40, 40, 40]                                 [Sun, Mon, Tues, Wed, Thur]
1  2020-12-24  2020-12-29                      [52, 52, 52, 52, 53, 53]                            [Thur, Fri, Sat, Sun, Mon, Tues]
2  2020-12-24  2021-01-03  [52, 52, 52, 52, 53, 53, 53, 53, 53, 53, 53]  [Thur, Fri, Sat, Sun, Mon, Tues, Wed, Thur, Fri, Sat, Sun]
3  2020-12-28  2021-01-03                  [53, 53, 53, 53, 53, 53, 53]                       [Mon, Tues, Wed, Thur, Fri, Sat, Sun]

Answer 2

你可以这样做：

def f(x):
    didx = pd.date_range(x['Date_1'], x['Date_2'])
    return pd.Series([didx.isocalendar().week.values, 
                      didx.strftime('%a').values], 
                      index=['Weeks', 'Days'])

df[['Weeks', 'Days']] = df.apply(f, axis=1)
  

输出：

      Date_1     Date_2                                         Weeks                                               Days
0 2020-09-27 2020-10-01                          [39, 40, 40, 40, 40]                          [Sun, Mon, Tue, Wed, Thu]
1 2020-12-24 2020-12-29                      [52, 52, 52, 52, 53, 53]                     [Thu, Fri, Sat, Sun, Mon, Tue]
2 2020-12-24 2021-01-03  [52, 52, 52, 52, 53, 53, 53, 53, 53, 53, 53]  [Thu, Fri, Sat, Sun, Mon, Tue, Wed, Thu, Fri, ...
3 2020-12-28 2021-01-03                  [53, 53, 53, 53, 53, 53, 53]                [Mon, Tue, Wed, Thu, Fri, Sat, Sun]