BUG: pyarrow string not saved for index column when writing parquet

[qltwis]

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
pd.DataFrame({
 "a":["aiwj","odko"],
 "b":["adwl","gfoe"],
 "c":["okdp","mbna"]
}).astype("string[pyarrow]").set_index("a").to_parquet("test.parq")

pd.read_parquet("test.parq").index.dtype

Issue Description

The dtype information for the index column is lost when saving as parquet.

Expected Behavior

pd.read_parquet("test.parq").index.dtype should return string[python] but instead gives dtype('O')

Installed Versions

INSTALLED VERSIONS ------------------ commit : 37ea63d python : 3.10.6.final.0 python-bits : 64 OS : Linux OS-release : 5.19.0-41-generic Version : #42~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue Apr 18 17:40:00 UTC 2 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 2.0.1
numpy : 1.22.4
pytz : 2023.3
dateutil : 2.8.2
setuptools : 65.3.0
pip : 22.2.2
Cython : None
pytest : 7.1.2
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : 2.9.3
jinja2 : 3.1.2
IPython : 8.4.0
pandas_datareader: None
bs4 : 4.12.2
bottleneck : None
brotli :
fastparquet : None
fsspec : 2023.4.0
gcsfs : None
matplotlib : 3.7.1
numba : 0.55.2
numexpr : None
odfpy : None
openpyxl : 3.0.10
pandas_gbq : None
pyarrow : 10.0.1
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : 1.10.1
snappy : None
sqlalchemy : 1.4.45
tables : None
tabulate : None
xarray : None
xlrd : None
zstandard : None
tzdata : 2023.3
qtpy : None
pyqt5 : None

[mroeschke]

Thanks for the report. This will be fixed in pyarrow 12 so closing as it will be fixed upstream: apache/arrow#34445

[qltwis]

Thanks for the report. This will be fixed in pyarrow 12 so closing as it will be fixed upstream: apache/arrow#34445

Thank you for the info. With pyarrow 12 pd.read_parquet("test.parq", dtype_backend="pyarrow").index.dtype produces the correct index dtype.

However, with the numpy backend the index type is still read as object. The other columns are correctly identified as string[python]. I imagine that's not intended behavior, right?

[qltwis]

Unfortunately this bug persists despite the fix in pyarrow 12.

[mroeschke]

This is unfortunately the intended behavior as long as pandas default string type is object. parquet does not encode how the string type implementation from pandas.

This can change in pandas 3.0 when the default string implementation is pyarrow if installed as a dependency. Otherwise one will need to specify the dtype_backend argument to recover the implementation so closing