使用熊猫从现有列中添加新日期列

Question

我有一个日期和星期的数据框，如下所示：

weeks = ["2016 W08","2016 W09","2016 W10","2016 W11"]
date = ["2-22-16", "2-29-16", "3-7-16", "3-14-16"]
df = pd.DataFrame(
{'week': weeks,
 'date': date,
 })
df
          date      week
0   2016-02-22  2016 W08
1   2016-02-29  2016 W09
2   2016-03-07  2016 W10
3   2016-03-14  2016 W11

我正在尝试非常简单地创建一个新列end，它比date 的值晚7 天。

我完全不喜欢 for 循环，并且想念如何循环遍历每个索引。

import datetime
for row in df['date']:
    start_date = df[row]
    df['end_date'] = start_date + datetime.timedelta(days = 7)

我知道我已经接近答案但错过了！

Answer 1

from datetime import timedelta
df.date=pd.to_datetime(df.date)
df.date+timedelta(days=7)

Out[691]: 
0   2016-02-29
1   2016-03-07
2   2016-03-14
3   2016-03-21
Name: date, dtype: datetime64[ns]

如果需要for循环：

import datetime
for i,row in df.iterrows():
    df.loc[i,'new']= row['date'] + datetime.timedelta(days = 7)
df
Out[698]: 
        date      week        new
0 2016-02-22  2016 W08 2016-02-29
1 2016-02-29  2016 W09 2016-03-07
2 2016-03-07  2016 W10 2016-03-14
3 2016-03-14  2016 W11 2016-03-21

编辑：为什么你的循环不起作用

for row in df['date']:
    start_date = row
    print(start_date + datetime.timedelta(days=7))

row在这里返回日期的值，不能用自己的值调用df的值。

Answer 2

试试pd.DateOffset:

import pandas as pd
weeks = ["2016 W08","2016 W09","2016 W10","2016 W11"]
date = ["2-22-16", "2-29-16", "3-7-16", "3-14-16"]
df = pd.DataFrame(
{'week': weeks,
 'date': date,
 })
df['date'] = pd.to_datetime(df['date']) #convert date column to datetime format
df['end_date'] = pd.DatetimeIndex(df['date']) + pd.DateOffset(7) #create new column end_date which is 7 days later
df

df 将是：

    date        week        end_date
0   2016-02-22  2016 W08    2016-02-29
1   2016-02-29  2016 W09    2016-03-07
2   2016-03-07  2016 W10    2016-03-14
3   2016-03-14  2016 W11    2016-03-21