使用开始日期和结束日期创建一个新列：Pandas

Question

我有一个如下所示的 pandas 数据框（示例），我想创建一个带有额外列 'NewDate' 的新 table，它将查看 StartDate 并显示最后日期每个 ID 的开始日期和随后的每个月的最后一个日期，直到结束日期，如果我的 ID 的结束日期为 Null，则该系列将在当月的最后一个日期即 2022 年 5 月停止。

ID     StartDate      EndDate            
100     1/01/2022     26/04/2022    
101    20/04/2022     Null
102     1/01/2022     27/02/2022          
....

我的预期输出：

ID  StartDate       EndDate          NewDate
100     1/01/2022    26/04/2022      31/01/2022
100     1/01/2022    26/04/2022      28/02/2022
100     1/01/2022    26/04/2022      31/03/2022
100     1/01/2022    26/04/2022      30/04/2022
101     20/04/2022    Null           30/04/2022
101     20/04/2022    Null           31/05/2022
102     1/01/2022    27/02/2022      31/01/2022
102     1/01/2022    27/02/2022      28/02/2022
...

Answer 1

试试这个

# convert each date column to datetime
df['StartDate'] = pd.to_datetime(df['StartDate'])
df['EndDate'] = pd.to_datetime(df['EndDate'])
# create date ranges for each row
f = lambda s,e: pd.date_range(s, e+pd.DateOffset(months=1), freq='M')
df['NewDate'] = [f(s,e) if e==e else f(s,pd.datetime.now()) for s, e in zip(df['StartDate'], df['EndDate'])]
# explode the new column
df = df.explode('NewDate')
print(df)

    ID  StartDate    EndDate    NewDate
0  100 2022-01-01 2022-04-26 2022-01-31
0  100 2022-01-01 2022-04-26 2022-02-28
0  100 2022-01-01 2022-04-26 2022-03-31
0  100 2022-01-01 2022-04-26 2022-04-30
1  101 2022-04-20        NaT 2022-04-30
1  101 2022-04-20        NaT 2022-05-31
2  102 2022-01-01 2022-02-27 2022-01-31
2  102 2022-01-01 2022-02-27 2022-02-28

Answer 2

您首先需要 trim 您的日期，然后我们使用 pd.date_range 和 explode 列创建日期范围

s1 = pd.to_datetime(df.StartDate, format = '%d/%m/%Y')
s2 = pd.to_datetime(df.EndDate, format = '%d/%m/%Y', errors = 'coerce') + pd.offsets.MonthEnd(0)
     
s2 = s2.fillna(s1 + pd.offsets.MonthEnd(2))

df['new'] = [pd.date_range(x, y , freq= 'M',closed = 'left') for x , y in zip(df.StartDate, s+pd.offsets.MonthEnd(1))]
out = df.explode('new')

out
Out[206]: 
    ID   StartDate     EndDate        new
0  100   1/01/2022  26/04/2022 2022-01-31
0  100   1/01/2022  26/04/2022 2022-02-28
0  100   1/01/2022  26/04/2022 2022-03-31
0  100   1/01/2022  26/04/2022 2022-04-30
1  101  20/04/2022        Null 2022-04-30
1  101  20/04/2022        Null 2022-05-31
2  102   1/01/2022  27/02/2022 2022-01-31
2  102   1/01/2022  27/02/2022 2022-02-28

已更新

s2 = s2.fillna(s1 + pd.offsets.MonthEnd(1))  
df['new_date'] = [pd.date_range(x, y , freq= 'M',closed = 'left') for x , y in zip(df.start_date, s2+pd.offsets.MonthEnd(1))] 
output = df.explode('new_date')

使用开始日期和结束日期创建一个新列：Pandas

Using Start Date and End Date create a a new column : Pandas

python

datetime

pandas