Feature Type

[ ] Adding new functionality to pandas
[X] Changing existing functionality in pandas
[ ] Removing existing functionality in pandas

Problem Description

For a data series with datetime index... I wish there is an easy way to make the very first element to be on the left side of the window when rolling starts.

As of pandas 2.1.4 I can align a rolling window so the very first element of my series is on the right side of the window when the rolling starts (i.e. center=False). I can also align the verify first element to be in the middle of the window when rolling starts (i.e. center=True)

But there is no easy way to make the very first element to be on the left side of the window when rolling starts.

For integer window I can see that I can use FixedForwardWindowIndexer as shown in the following page https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.api.indexers.FixedForwardWindowIndexer.html#pandas.api.indexers.FixedForwardWindowIndexer

But that doesn't seem to work for when the index is datetime.

Feature Description

import pandas as pd

# create series
s = pd.Series(range(5), index=pd.date_range('2020-01-01', periods=5, freq='1D'))

# create a function to call for each window
def roll_apply(window):
    # return first element of current window
    return window[:1].iloc[0]

# if center=False, the very first window will have the first element of my series on its right most side
right = s.rolling('4D', center=False).apply(roll_apply)

# if center=False, the very first window will have the first element of my series on its center
center = s.rolling('4D', center=True).apply(roll_apply)

# there is no easy way to have the very first window have the first element of my series on its left
# ???


print("********* original *********")
print(s)

print("********* right aligned *********")
print(right)

print("********* center aligned *********")
print(center)

print("********* left aligned *********")
print("Not Implemented Yet!")

output

********* original *********
2020-01-01    0
2020-01-02    1
2020-01-03    2
2020-01-04    3
2020-01-05    4
Freq: D, dtype: int64
********* right aligned *********
2020-01-01    0.0
2020-01-02    0.0
2020-01-03    0.0
2020-01-04    0.0
2020-01-05    1.0
Freq: D, dtype: float64
********* center aligned *********
2020-01-01    0.0
2020-01-02    0.0
2020-01-03    1.0
2020-01-04    2.0
2020-01-05    3.0
Freq: D, dtype: float64
********* left aligned *********
Not Implemented Yet!

Alternative Solutions