[SPARK-38940][PYTHON] Test Series' anchor frame for in-place updates on Series

[xinrong-meng]

What changes were proposed in this pull request?

Test Series' anchor DataFrame for in-place updates on Series.

Why are the changes needed?

In pandas, some in-place updates on Series change its anchor frame (e.g. Series.clip), while others don't.
We should test that to ensure parity with pandas.

Example on changed anchor DataFrame

>>> df = pd.DataFrame({'x': [1, 2, 3]})
>>> ser = df.x

>>> ser.clip(2, 2, inplace=True)
>>> ser
0    2
1    2
2    2
Name: x, dtype: int64
>>> df
   x
0  2
1  2
2  2

Example on unchanged anchor DataFrame

>>> df = pd.DataFrame({'x': [1, 2, 3]})
>>> ser = df.x

>>> ser.drop(1, inplace=True)
>>> ser
0    1
2    3
Name: x, dtype: int64
>>> df
   x
0  1
1  2
2  3

Does this PR introduce any user-facing change?

No. Test only.

How was this patch tested?

Unit tests.

[xinrong-meng]

Note that Series.rename_axis has a bug so we may not test Series' anchor frame in this PR. The fix and the test will be added together in a followup PR.

[xinrong-meng]

@ueshin @HyukjinKwon @itholic Thank you!

Please refer to https://github.com/xinrong-databricks/spark/actions/runs/2186563996 for test results.

[HyukjinKwon]

Merged to master.