BUG: DatetimeIndex.tz_convert() has incorrcect .freq attribute

[rwijtvliet]

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

# DST-transition. Frequency is days, because all timestamps are at same time - even if not all days span 24h. So far, so good.
i = pd.date_range('2020-03-27 06:00', freq='D', periods=5, tz='Europe/Berlin')
# DatetimeIndex(['2020-03-27 06:00:00+01:00', '2020-03-28 06:00:00+01:00',
#                '2020-03-29 06:00:00+02:00', '2020-03-30 06:00:00+02:00',
#                '2020-03-31 06:00:00+02:00'],
#               dtype='datetime64[ns, Europe/Berlin]', freq='D')

# Converting to other timezone gives timestamps with distinct times. This index should no longer have the daily frequency. However: it is initially given this frequency nonetheless (=Bug).
i2 = i.tz_convert('UTC')
# DatetimeIndex(['2020-03-27 05:00:00+00:00', '2020-03-28 05:00:00+00:00',
#                '2020-03-29 04:00:00+00:00', '2020-03-30 04:00:00+00:00',
#                '2020-03-31 04:00:00+00:00'],
#               dtype='datetime64[ns, UTC]', freq='D')

# We can confirm that this is incorrect, by manually trying to set the `.freq` attribute:
i2.freq = i2.freq #ValueError: Inferred frequency None from passed values does not conform to passed frequency D

Issue Description

See inline comments.

Expected Behavior

i2 should not have a frequency attribute set.

Installed Versions

INSTALLED VERSIONS

commit : 91111fd
python : 3.9.13.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19045
machine : AMD64
processor : Intel64 Family 6 Model 165 Stepping 2, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : de_DE.cp1252

pandas : 1.5.1
numpy : 1.23.4
pytz : 2022.6
dateutil : 2.8.2
setuptools : 63.4.1
pip : 22.2.2
Cython : None
pytest : 7.1.3
hypothesis : None
sphinx : 5.3.0
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.1.2
IPython : 8.5.0
pandas_datareader: None
bs4 : 4.11.1
bottleneck : None
brotli : None
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : 3.6.2
numba : None
numexpr : None
odfpy : None
openpyxl : 3.0.10
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

[Detoro]

take

[keelancosgrove]

hi @rwijtvliet , the frequency attribute was set as part of this issue #25241 - why do you think it should not be set in tz_convert?

[rwijtvliet]

Hey @keelancosgrove, in the example above, i2 contains timestamps that are on consequtive days, but not at the same time on those days - which is be necessary for them to be "1 day" apart (in my opinion). Therefore, in this case the frequency for i2 should no longer be "day".