BUG: freq attribute is None after set_index

[antvig]

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

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

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

df = pd.DataFrame({"datetime": pd.date_range("1/1/1", periods=20, freq="H", tz="UTC")})

df.set_index("datetime", inplace=True)

df.index.freq --> `None`

Problem description

This behavior is new in 1.3.0 and 1.3.1.

In 1.2.5, the output was, as expected, H

Expected Output

'H'

Output of pd.show_versions()

INSTALLED VERSIONS

commit : c7f7443
python : 3.8.10.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 158 Stepping 13, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : fr_FR.cp1252
pandas : 1.3.1
numpy : 1.21.1
pytz : 2021.1
dateutil : 2.8.2
pip : 21.1.3
setuptools : 57.4.0
Cython : None
pytest : 6.2.4
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.7.0
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : None

[simonjayhawkins]

Thanks @antvig for the report.

This behavior is new in 1.3.0 and 1.3.1.

first bad commit: [6b6d8fd] REF: Back DatetimeTZBlock with sometimes-2D DTA (#41082) cc @jbrockmendel

[jbrockmendel]

This behavior matches what happens with tznaive DTI (which i think theres an issue about), is unavoidable

[simonjayhawkins]

changing milestone to 1.3.5

[jbrockmendel]

@simonjayhawkins there's essentially no prospect for this to change for 1.3.5

[simonjayhawkins]

@simonjayhawkins there's essentially no prospect for this to change for 1.3.5

Can you elaborate. Is the new behavior considered correct? If not, why can't this be fixed.

[jreback]

good to fix but not needed for 1.3.x

[jbrockmendel]

Can you elaborate. Is the new behavior considered correct? If not, why can't this be fixed.

The changed behavior is driven by the fact that when you do reset_index with a DTI/TDI, it is unwrapped to DTA/TDA to become a column and reshaped to 2D. 2D DTA/TDA never have non-None freq. We could "cheat" here since it is single-column and retain the freq, but that would only be until 2.0 when DTA/TDA will lost the freq attr entirely (xref #31218)

[simonjayhawkins]

moving off 1.3.x milestone