BUG: maybe_convert_objects fails for byte series

[lbittarello]

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.Series([b"a"], dtype="bytes").infer_objects()

Issue Description

The nightly version of maybe_convert_objects has been throwing an error for byte series for the past couple of days. Not sure if the recent refactoring (#49588) has something to do with it.

Expected Behavior

The code runs through and returns the unmodified series with the latest release.

Installed Versions

INSTALLED VERSIONS
------------------
commit           : 6ba52161147f2543d6cd4194958b0071fec6d1dc
python           : 3.11.0.final.0
python-bits      : 64
OS               : Windows
OS-release       : 10
Version          : 10.0.22621
machine          : AMD64
processor        : AMD64 Family 23 Model 96 Stepping 1, AuthenticAMD
byteorder        : little
LC_ALL           : None
LANG             : None
LOCALE           : English_United Kingdom.1252

pandas           : 2.0.0.dev0+642.g6ba5216114
numpy            : 1.23.4
pytz             : 2022.6
dateutil         : 2.8.2
setuptools       : 65.5.1
pip              : 22.3.1
Cython           : None
pytest           : None
hypothesis       : None
sphinx           : None
blosc            : None
feather          : None
xlsxwriter       : None
lxml.etree       : None
html5lib         : None
pymysql          : None
psycopg2         : None
jinja2           : None
IPython          : 8.6.0
pandas_datareader: None
bs4              : None
bottleneck       : None
brotli           : 
fastparquet      : None
fsspec           : None
gcsfs            : None
matplotlib       : None
numba            : None
numexpr          : None
odfpy            : None
openpyxl         : None
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
zstandard        : None
tzdata           : None
qtpy             : None
pyqt5            : None
c:\Programs\Python\envs\pandas_nightly\Lib\site-packages\_distutils_hack\__init__.py:33: UserWarning: Setuptools is replacing distutils.
  warnings.warn("Setuptools is replacing distutils.")

[phofl]

Hi, thanks for your report. This was caused by #49590

cc @jbrockmendel Was this intended?

[jbrockmendel]

no that was not intended

[jbrockmendel]

though i think something may be fishy with dtype="bytes", id expect that to cast to object dtype

[phofl]

This is part of the wider problem that we sometimes end up with numpy str or bytes dtype and cast in some other cases.

Would you be ok with reverting the try/except removal till we can come up with a general solution?

[jbrockmendel]

Would you be ok with reverting the try/except removal till we can come up with a general solution?

I'd suggest adding a check in ObjectBlock.convert for self.dtype == object

[phofl]

Just that I understand you: You would cast self.values to object if this check fails?

This would not get rid of the problem that we sometimes end up with bytes dtype in the Series constructor

[jbrockmendel]

no. if its self.dtype!=object then there is no need to do inference