BUG: Bad shift_forward ceiling around DST change

[epizut]

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

just_before_dst = pd.to_datetime('2023-03-26 00:45:00').tz_localize('Europe/London')
just_before_dst.ceil('30min', nonexistent='shift_forward')
# Result:    2023-03-26 03:00:00
# Excpected: 2023-03-26 02:00:00

Issue Description

On March 26th 2023 London will switch to its DST time and goes directly from 00:59.99 to 02:00:00 skipping 1h of local time.
Unfortunately, ceiling 00:45:00 to its upper hour results into 03:00:00.

Expected Behavior

2023-03-26 02:00:00

Installed Versions

INSTALLED VERSIONS ------------------ commit : 2e218d1 python : 3.9.12.final.0 python-bits : 64 OS : Windows OS-release : 10 Version : 10.0.19042 machine : AMD64 processor : Intel64 Family 6 Model 85 Stepping 4, GenuineIntel byteorder : little LC_ALL : None LANG : None LOCALE : English_United States.1252

pandas : 1.5.3
numpy : 1.23.5
pytz : 2022.7.1
dateutil : 2.8.2
setuptools : 65.3.0
pip : 22.2.2
Cython : 0.29.30
pytest : 7.2.1
hypothesis : None
sphinx : 4.5.0
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.9.0
html5lib : None
pymysql : None
psycopg2 : 2.9.3
jinja2 : 3.1.1
IPython : 8.2.0
pandas_datareader: None
bs4 : 4.11.1
bottleneck : None
brotli : None
fastparquet : None
fsspec : 2022.3.0
gcsfs : None
matplotlib : 3.5.1
numba : 0.56.4
numexpr : None
odfpy : None
openpyxl : 3.0.9
pandas_gbq : None
pyarrow : 7.0.0
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : 1.8.1
snappy : None
sqlalchemy : 1.4.41
tables : None
tabulate : 0.8.9
xarray : 2022.3.0
xlrd : 2.0.1
xlwt : None
zstandard : None
tzdata : None

[ivyrxr]

take

[ivyrxr]

The result "2023-03-26 03:00:00" is expected because when you shift 00:45:00 forwards by 30 mins, it becomes 01:15:00 which is actually 02:15:00 in London DST time. Ceiling it will return 03:00:00. Hope it helps or please correct me if I am wrong.

[epizut]

I don't get it, even if we forget the DST:
pd.to_datetime('2023-03-26 02:15:00').ceil('30min') != pd.to_datetime('2023-03-26 03:00:00')

If the correct answer is 2023-03-26 03:00:00 London then it would be 1h 15minutes after the original time 2023-03-26 00:45:00 London, but I don't expect ceil('30min') to shift by more than 29:59.999999

[MarcoGorelli]

thanks for the report - I think I would also have expected 2023-03-26 02:00:00

[ivyrxr]

Thank you both for the clarification. After I read more closely to the documentation, I think you are right that the expected result should be 2023-03-26 02:00:00. I can help investigate the code for a fix.

[MarcoGorelli]

that would be great, thank you! feel free to ping if anything trips you up