BUG: inplace behavior changes after 1.4.x

[Yikun]

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

pdf = pd.DataFrame({"A": range(1, 3), "B": range(4, 6), "C": range(7, 9)})
pser = pdf.A
pdf.eval("A = B + C", inplace=True)

Issue Description

Before 1.4.0:

>>> pdf.A
0    11
1    13
Name: A, dtype: int64
>>> pser
0    11
1    13
Name: A, dtype: int64

After 1.4.0:

>>> pdf.A
0    11
1    13
Name: A, dtype: int64
>>> pser   # <== behavior changed here
0    1
1    2
Name: A, dtype: int64

Expected Behavior

>>> pdf.A
0    11
1    13
Name: A, dtype: int64
>>> pser
0    11
1    13
Name: A, dtype: int64

Installed Versions

1.4.x+

[Yikun]

pdf = pd.DataFrame({"A": range(1, 3), "B": range(4, 6), "C": range(7, 9)})
pser = pdf.A
pdf["A"] = pdf["A"] + 10
pser
pdf.A

Here is a simple test

[Yikun]

first change commit: 03dd698

cc @jbrockmendel

[phofl]

Hi, thanks for your report.

This was wrong before 1.4.0 and works as expected now. This was also announced in the release notes. Setitem should always make a copy and never write into the existing array.