BUG: unexpected behavior in read_csv with pyarrow engine and dtype string

[alejoforero]

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

df = pd.DataFrame({'var1': ['44794724', None]})
df.to_csv('sample.csv', index=False)
df.head

df1 = pd.read_csv('sample.csv', dtype='string[pyarrow]',dtype_backend='pyarrow')
df2 = pd.read_csv('sample.csv', dtype='string[pyarrow]',dtype_backend='pyarrow',engine="pyarrow")

print(df1['var1'].astype(str).str.endswith('.0').sum())
print(df2['var1'].astype(str).str.endswith('.0').sum())

Issue Description

I've encountered an issue with read_csv when using the pyarrow engine. Specifically, when a numeric column with missing values is imported as a string type, each number in the column has ".0" appended to it. This issue occurs when setting the dtype to 'string' or 'string[pyarrow]'. The issue is not observed if at least one of the data rows contains a string, or if the dtype is set to 'object'.

Expected Behavior

Expected the same behavior turning on and off the pyarrow engine. The numeric values should be imported as strings without alteration.

Installed Versions

pandas '2.2.0' and '3.0.0.dev0+171.gd928a5cc22'
pyarrow=14.0.2
python=3.11.7
OS: win11-22H2

[vinodseshadri]

take

[vinodseshadri]

Able to reproduce the issue with both 2.2.0 and 1.5.3 Triaging the issue further

[vinodseshadri]

For cases where dtype_backend is not pyarrow, the fix is to modify the table.to_pandas() here to take in a new parameter and change to frame = table.to_pandas(integer_object_nulls=True).

This fix does not seems to be fixing the case where dtype_backend='pyarrow' . Need to explore further