BUG: first() and last() compare times to first/last row according to position rather than first/last row according to index time

[mvashishtha]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd

i = pd.date_range('2018-04-09', periods=4, freq='2D')
ts = pd.DataFrame({'A': [1, 2, 3, 4]}, index=i)
# Prints the times from the first 3 days (so first two rows)
print(ts.first('3D'))
# Prints empty dataframe
print(ts.iloc[::-1].first('3D'))

# Prints the times from the last 3 days (so last two rows)
print(ts.last('3D'))
# Prints empty dataframe
print(ts.iloc[::-1].last('3D'))

Issue Description

When I have ts ordered by increasing date, first('3D') selects the two rows that are within 3 days, inclusive, of the earliest date, 2018-04-09. If I call first('3D') on ts reversed, I get no results. Same applies mutatis mutandis to last.

Expected Behavior

first('3D') should give me the two rows that are within 3 days, inclusive, of the earliest date, 2018-04-09, so the rows with 2018-04-09 and 2018-04-11. The description of first() suggests that it gives the first dates relative to the temporal first date in the index. Same applies mutatis mutandis to last.

Installed Versions

/Users/maheshvashishtha/opt/anaconda3/envs/pandas-dev-alt/lib/python3.8/site-packages/_distutils_hack/init.py:33: UserWarning: Setuptools is replacing distutils.
warnings.warn("Setuptools is replacing distutils.")

INSTALLED VERSIONS

commit : 2e218d1
python : 3.8.16.final.0
python-bits : 64
OS : Darwin
OS-release : 21.5.0
Version : Darwin Kernel Version 21.5.0: Tue Apr 26 21:08:22 PDT 2022; root:xnu-8020.121.3~4/RELEASE_X86_64
machine : x86_64
processor : i386
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.5.3
numpy : 1.24.1
pytz : 2022.7.1
dateutil : 2.8.2
setuptools : 65.6.3
pip : 22.3.1
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : 8.8.0
pandas_datareader: None
bs4 : None
bottleneck : None
brotli : None
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : None
snappy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
zstandard : None
tzdata : None

[dicristina]

df.first [df.last] takes the first [last] item of the index and adds [subtracts] the offset to get the end [start] timestamp then it searches the index (with Index.searchsorted) for the offset of the end [start] value and finally slices the DataFrame. This works only if the index is sorted in ascending order but the documentation does not state this.

[dicristina]

df.first gives different result for MonthBegin and MonthEnd:

In [1]: import pandas as pd

In [2]: s = pd.Series(1, index=pd.bdate_range("2010-03-31", periods=100))

In [3]: print(s.first("1M"))   # As expected
2010-03-31    1
Freq: B, dtype: int64

In [4]: print(s.first("1MS"))
2010-03-31    1
2010-04-01    1
Freq: B, dtype: int64

I think it would be best to rewrite it so that the results always match the groupings given by df.resample.

[MarcoGorelli]

thanks for the report

this function is being deprecated, see #45908 (comment), so let's close for now