BUG: `USFederalHolidayCalendar.holidays` inconsistently returns Index and not DateTimeIndex

[austinjpaul]

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 datetime
>>> from pandas.tseries.holiday import USFederalHolidayCalendar
>>> cal = USFederalHolidayCalendar()
>>> cal.holidays(start=datetime.date(2018,8,1), end=datetime.date(2018,8,31))
Index([], dtype='object')

Issue Description

The returned result should be type DatetimeIndex.
Non-empty results are indeed of type DatetimeIndex. Furthermore, certain other empty results from other ranges (seen below in expected behavior) also have type DatetimeIndex.
The return type is inconsistent.

Expected Behavior

>>> cal.holidays(start=datetime.date(2020,8,1), end=datetime.date(2020,8,31))
DatetimeIndex([], dtype='datetime64[ns]', freq=None)
>>> cal.holidays(start=datetime.date(2019,8,1), end=datetime.date(2019,8,31))
DatetimeIndex([], dtype='datetime64[ns]', freq=None)

Installed Versions

INSTALLED VERSIONS ------------------ commit : 87cfe4e python : 3.10.6.final.0 python-bits : 64 OS : Linux OS-release : 5.10.102.1-microsoft-standard-WSL2 Version : #1 SMP Wed Mar 2 00:30:59 UTC 2022 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : C.UTF-8 LOCALE : en_US.UTF-8

pandas : 1.5.0
numpy : 1.23.3
pytz : 2022.4
dateutil : 2.8.2
setuptools : 65.4.1
pip : 22.2.2

[roadswitcher]

Same behavior seen for March (the other month with no US Federal holidays), but interestingly the behavior is consistent for 2019.

>>> from pandas.tseries.holiday import USFederalHolidayCalendar
>>> cal = USFederalHolidayCalendar()
>>> cal.holidays(start=datetime.date(2018,8,1), end=datetime.date(2018,8,31))
Index([], dtype='object')
>>> cal.holidays(start=datetime.date(2018,3,1), end=datetime.date(2018,3,31))
Index([], dtype='object')
>>> cal.holidays(start=datetime.date(2019,8,1), end=datetime.date(2019,8,31))
DatetimeIndex([], dtype='datetime64[ns]', freq=None)
>>> cal.holidays(start=datetime.date(2019,3,1), end=datetime.date(2019,3,31))
DatetimeIndex([], dtype='datetime64[ns]', freq=None)

[roadswitcher]

Yeah, I missed some strange behavior that made my PR irrelevant.

Documenting for the record here:

• submitters original issue can be replicated by picking any initial pair of dates that do not contain a holiday between them, and get an Index object.
• You can then call against the same calendar object with a date range that contains at least one holiday, and get the DatetimeIndex submitter expected
• Repeat the original request with a range of dates that do not contain a holiday, and you now get a DatetimeIndex.

Example:

In [3]: cal=USFederalHolidayCalendar()

In [4]: cal.holidays(start="2018-08-01", end="2018-08-31")
Out[4]: Index([], dtype='object')

In [5]: cal.holidays(start="2018-07-05", end="2018-08-31")
Out[5]: Index([], dtype='object')

In [6]: cal.holidays(start=datetime.date(2018,8,1), end=datetime.date(2018,8,31))
Out[6]: Index([], dtype='object')

In [7]: cal.holidays(start=datetime.date(2018,7,1), end=datetime.date(2018,8,31))
Out[7]: DatetimeIndex(['2018-07-04'], dtype='datetime64[ns]', freq=None)

In [8]: cal.holidays(start=datetime.date(2018,8,1), end=datetime.date(2018,8,31))
Out[8]: DatetimeIndex([], dtype='datetime64[ns]', freq=None)

[roadswitcher]

take