BUG: invalid result of reindex on columns after unstack with Period data

[a-rvn]

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

series1 = pd.DataFrame(
   [(0, "s2", pd.Period(2022)), (0, "s1", pd.Period(2021))],
   columns=["A", "B", "C"]
).set_index(["A", "B"])["C"]

series2 = series1.astype(str)

print(series1.unstack("B").reindex(["s2"], axis=1))
print(series2.unstack("B").reindex(["s2"], axis=1))

Issue Description

The example code prints

B  s2
A      
0  2021

B  s2
A  
0  2022

The result with pd.Period data is the incorrect 2021, but with str data it's the correct 2022.

This only occurs with certain index values. When replacing "s1" with "s3", the effect disappears and the result is 2022 in both cases.

Expected Behavior

Expect the result for both pd.Period and str data to be 2022:

B  s2
A  
0  2022

B  s2
A  
0  2022

(actually observed with older Pandas 2.0.3)

Installed Versions

INSTALLED VERSIONS

commit : 0691c5c
python : 3.11.10
python-bits : 64
OS : Linux
OS-release : 6.2.16
Version : #1-NixOS SMP PREEMPT_DYNAMIC Tue Jan 1 00:00:00 UTC 1980
machine : x86_64
processor :
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8
pandas : 2.2.3
numpy : 2.2.3
pytz : 2025.1
dateutil : 2.9.0.post0
pip : 24.0
Cython : None
sphinx : None
IPython : None
adbc-driver-postgresql: None
adbc-driver-sqlite : None
bs4 : None
blosc : None
bottleneck : None
dataframe-api-compat : None
fastparquet : None
fsspec : None
html5lib : None
hypothesis : None
gcsfs : None
jinja2 : None
lxml.etree : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
psycopg2 : None
pymysql : None
pyarrow : None
pyreadstat : None
pytest : None
python-calamine : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlsxwriter : None
zstandard : None
tzdata : 2025.1
qtpy : None
pyqt5 : None

[pedromfdiogo]

take

[rhshadrach]

Thanks for the report! Confirmed on main. There has to be some internal state that gets set wrong. Constructing the unstacked result directly:

series1 = pd.DataFrame(
   [(0, "s2", pd.Period(2022)), (0, "s1", pd.Period(2021))],
   columns=["A", "B", "C"]
).set_index(["A", "B"])["C"]
unstacked = series1.unstack("B")
df = pd.DataFrame({"s1": pd.Period(2021), "s2": pd.Period(2022)}, index=pd.Index([0], name="A"))
df.columns.name = "B"
tm.assert_frame_equal(unstacked, df)

does not exhibit the buggy behavior either when using reindex. Further investigations and PRs to fix are welcome!

[snitish]

@rhshadrach I noticed that PeriodDtype (along with DatetimeTZDtype) is set to support 2D values in the block:

pandas/pandas/core/dtypes/dtypes.py

Line 1034 in 56847c5

_supports_2d = True

but in method Block._validate_ndim we only check for DatetimeTZDtype and not PeriodDtype.

pandas/pandas/core/internals/blocks.py

Line 162 in 56847c5

dtype, DatetimeTZDtype

Adding PeriodDtype to the isinstance check seems to fix this issue, but wondering if we should use dtype._supports_2d instead of individually comparing with each known 2D dtype.