按自定义年度频率重新采样 [英] Resample by custom annual frequency

问题描述

我知道我可以从7月1日开始每年使用AS-JUL进行重新采样，但是如何在不同的日期进行采样?

I know I can use AS-JUL to resample annually starting in July 1st, but how can I do by a different date?

In [11]: df
Out[11]:
            value
date
2005-07-02      4
2005-09-20      7
2005-11-12      4
2005-12-31      0
2006-07-03      5
2006-06-25      6
2006-07-23      1
2006-09-28      8
2006-12-21      1
2006-12-27      9
2007-07-03      7
2007-09-21      0

In [12]: df.resample("AS-JUL").sum()
Out[12]:
            value
date
2005-07-01     26
2006-07-01     19
2007-07-01      7

我希望开始日期为7月5日，如何在此之前重新采样?

I want the start date to be the 5th July, how can I resample by that?

推荐答案

您可以减去4天，重新采样，然后再添加4天...但这真是太可怕了:

You could subtract 4 days, resample, then add the 4 days back... but this is kind of horrible:

In [11]: data2.index = data2.index - pd.offsets.Day(4)

In [12]: res = data2.resample("AS-JUL").sum()

In [13]: res.index = res.index + pd.offsets.Day(4)

In [14]: res
Out[14]:  
           value
date    
2004-07-05  4
2005-07-05  22
2006-07-05  26
2007-07-05  12

这篇关于按自定义年度频率重新采样的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！