在 Pandas 中创建自定义大小的日期时间系列容器

Question

I have multiple Pandas Series of datetime64 values that I want to bin into groups using arbitrary bin sizes.

I've found the Series.to_period() function which does exactly what I want except that I need more control over the chosen bin size. to_period allows me to bin by full years, months, days, etc. but I also want to bin by 5 years, 6 hours or 15 minutes. Using a syntax like 5Y , 6H or 15min works in other corners of Pandas but apparently not here.

s = pd.Series(["2020-02-01", "2020-02-02", "2020-02-03", "2020-02-04"], dtype="datetime64[ns]")

# Output as expected
s.dt.to_period("M").value_counts()
2020-02    4
Freq: M, dtype: int64

# Output as expected
s.dt.to_period("W").value_counts()
2020-01-27/2020-02-02    2
2020-02-03/2020-02-09    2
Freq: W-SUN, dtype: int64

# Output as expected
s.dt.to_period("D").value_counts()
2020-02-01    1
2020-02-02    1
2020-02-03    1
2020-02-04    1
Freq: D, dtype: int64

# Output unexpected (and wrong?)
s.dt.to_period("2D").value_counts()
2020-02-01    1
2020-02-02    1
2020-02-03    1
2020-02-04    1
Freq: 2D, dtype: int64

Answer 1

I believe that pd.Grouper is what you're looking for.

https://pandas.pydata.org/docs/reference/api/pandas.Grouper.html

It provides the flexibility of having multiple frequencies in addition to the standard ones: https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html#offset-aliases

From the documentation:

>>> start, end = '2000-10-01 23:30:00', '2000-10-02 00:30:00'
>>> rng = pd.date_range(start, end, freq='7min')
>>> ts = pd.Series(np.arange(len(rng)) * 3, index=rng)
>>> ts
2000-10-01 23:30:00     0
2000-10-01 23:37:00     3
2000-10-01 23:44:00     6
2000-10-01 23:51:00     9
2000-10-01 23:58:00    12
2000-10-02 00:05:00    15
2000-10-02 00:12:00    18
2000-10-02 00:19:00    21
2000-10-02 00:26:00    24
Freq: 7T, dtype: int64

>>> ts.groupby(pd.Grouper(freq='17min')).sum()
2000-10-01 23:14:00     0
2000-10-01 23:31:00     9
2000-10-01 23:48:00    21
2000-10-02 00:05:00    54
2000-10-02 00:22:00    24
Freq: 17T, dtype: int64

NOTE: If you'd like to .groupby a certain column then use the following syntax: df.groupby(pd.Grouper(key="my_col", freq="3M"))