BUG: Get deprecation warning from Series constructor

[rikardn]

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

ind1 = pd.Series([1])
ind2 = pd.Series([2])
ser = pd.Series([1.23], index=[ind1, ind2])
print(ser)

Issue Description

If trying to create a series with a multiindex from two series I get a deprecation warning with pandas 2.1.1

DeprecationWarning: Series._data is deprecated and will be removed in a future version. Use public APIs instead.

Since I am not using _data explicitly it must come from inside of pandas. The only place I have found where this could come from is in the is_all_arraylike function where there is a hasattr(val, "_data") in the cython code.

Expected Behavior

I expect no warning and at least in pandas 2.0.3 this was the case.

Installed Versions

Replace this line with the output of pd.show_versions()

[lithomas1]

pandas/pandas/_libs/lib.pyx

Line 869 in 2a08b05

util.is_array(val) or hasattr(val, "_data")):

is the offending line.

cc @jbrockmendel

[jbrockmendel]

The hasattr check is bad. Changing it to hasattr(type(val), "_data") might avoid the warning for now?

[paulreece]

I can confirm that the deprecation warning occurs on the main development branch.

>>> import pandas as pd
>>> ind1 = pd.Series([1])
>>> ind2 = pd.Series([2])
>>> ser = pd.Series([1.23], index=[ind1, ind2])
<stdin>:1: DeprecationWarning: Series._data is deprecated and will be removed in a future version. Use public APIs instead.