如何在 pandas 数据帧中创建 4 年周期的日数？

创建分组器以标记周期更改的行，计算 cumcount 以分配每个组的行号

s = df['CyclePhase'].diff().lt(0).cumsum()
df['Day'] = s.groupby(s).cumcount() + 1

---

              Dateindex   Price  CyclePhase  Day
0   1928-01-03 00:00:00   71.04           0    1
1   1928-01-04 00:00:00   70.88           0    2
2   1928-01-05 00:00:00   70.20           0    3
3   1928-01-06 00:00:00   70.64           0    4
4   1929-05-09 00:00:00  104.08           1    5
5   1929-05-10 00:00:00  105.36           1    6
6   1929-05-11 00:00:00  104.96           1    7
7   1929-05-13 00:00:00  102.56           1    8
8   1930-11-08 00:00:00   63.56           2    9
9   1930-11-10 00:00:00   62.16           2   10
10  1930-11-11 00:00:00   63.16           2   11
11  1931-12-29 00:00:00   31.84           3   12
12  1931-12-30 00:00:00   32.40           3   13
13  1931-12-31 00:00:00   32.48           3   14
14  1932-01-02 00:00:00   31.28           0    1
15  1932-01-04 00:00:00   30.24           0    2
16  1932-01-05 00:00:00   30.20           0    3

您可以按年份分组，如下所示：

df['Day'] = df.groupby(df['Dateindex'].dt.year // 4).cumcount().add(1)

如果您的数据从某个年份开始，而该年份不能被 整除，那么您可以执行以下操作：4

years = df['Dateindex'].dt.year
df['Day'] = df.groupby(years.sub(years.min()) // 4).cumcount().add(1)

注意：如果 ur 确实是索引，那么Dateindex

years = df.index.year