python获取前一天时间_Python从pandas date列获取一周的前一天（在我的示例中为星期四）...

最新推荐文章于 2024-06-07 08:49:34 发布

极萨学院冷哲

最新推荐文章于 2024-06-07 08:49:34 发布

阅读量1k

点赞数

文章标签： python获取前一天时间

本文链接：https://blog.csdn.net/weixin_28896255/article/details/111951257

版权

当使用Pandas的`pd.offsets.Week`时，遇到跨年的情况，无法正确获取前一周的星期四日期。解决方法是使用`Series.where`结合`Series.dt.weekday`，只在非星期四的日期上应用偏移量。这样可以确保在任何日期都能返回正确的前一个星期四。

摘要由CSDN通过智能技术生成

I have a pandas date column and I want to return the date for the previous Thursday (could be any day of week). I use pd.offsets.Week but I do not get the expected result when the year changes and the Week starts over. Here is my dataframe as 'd':

raw date Thursday week_start

0 2019-01-03 2018-12-27 2018-12-27

1 2019-01-03 2018-12-27 2018-12-27

2 2019-01-03 2018-12-27 2018-12-27

3 2019-01-02 2018-12-27 2018-12-27

4 2019-01-02 2018-12-27 2018-12-27

5 2019-01-02 2018-12-27 2018-12-27

6 2019-01-03 2019-01-03 2018-12-27

7 2019-01-03 2019-01-03 2018-12-27

8 2019-01-03 2019-01-03 2018-12-27

9 2019-01-03 2019-01-03 2018-12-27

10 2019-01-02 2018-12-27 2018-12-27

11 2019-01-02 2018-12-27 2018-12-27

12 2019-01-02 2018-12-27 2018-12-27

d['week_start'] = d['raw date'] - pd.offsets.Week(weekday=3)

I expected where d['week_start'] = 1/3/2019 to return 1/3/2019, not 12/27/2018. I suspect it is because the Week at 1/3/2019 is 0 so it returns the Thursday of that week. How can I get the previous Thursday's date regardless of a change in the year?

解决方案

You can use Series.where with Series.dt.weekday for change only not Thursday values:

rng = pd.date_range('2019-01-03', periods=20)

d = pd.DataFrame({'raw date': rng})

mask = d['raw date'].dt.weekday == 3

d['week_start'] = d['raw date'].where(mask, d['raw date'] - pd.offsets.Week(weekday=3))

print(d)

raw date week_start

0 2019-01-03 2019-01-03

1 2019-01-04 2019-01-03

2 2019-01-05 2019-01-03

3 2019-01-06 2019-01-03