[SPARK-56014][PS][TESTS] Fix to_numeric ignore test for pandas 3.0

[ueshin]

What changes were proposed in this pull request?

This PR updates pyspark.pandas.tests.test_namespace.NamespaceTests.test_to_numeric for the pandas 3.0 behavior of to_numeric(..., errors="ignore") with non-Series inputs.

In this code path, ps.to_numeric delegates to pd.to_numeric for non-Series inputs. The existing test assumed that errors="ignore" returns the original input, but pandas 3.0 now raises ValueError("invalid error value specified") instead.

This patch makes the test follow the pandas version in use:

• for pandas < 3.0.0, keep the existing equality check
• for pandas >= 3.0.0, assert the ValueError

No implementation behavior is changed.

Why are the changes needed?

The current test fails under the pandas 3.0 test environment because its expectation no longer matches upstream pandas behavior.

Since pandas-on-Spark delegates this non-Series case to pandas, the test should reflect the version-specific pandas behavior rather than hard-coding the pre-3.0 result.

Does this PR introduce any user-facing change?

Yes, it will behave more like pandas 3.

How was this patch tested?

Updated the related test.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: OpenAI Codex (GPT-5)

[ueshin]

cc @gaogaotiantian @HyukjinKwon @zhengruifeng

[dongjoon-hyun]

Since the default value was raise, this looks okay to me.

BTW, do you think we need to re-fresh the doc, @ueshin ? The docs says ignore doesn't work for pandas-on-Spark Series args.

spark/python/pyspark/pandas/namespace.py

Lines 3595 to 3608 in 7eef6f7

	def to_numeric(arg, errors="raise"):
	"""
	Convert argument to a numeric type.
	Parameters
	----------
	arg : scalar, list, tuple, 1-d array, or Series
	Argument to be converted.
	errors : {'raise', 'coerce'}, default 'raise'
	* If 'coerce', then invalid parsing will be set as NaN.
	* If 'raise', then invalid parsing will raise an exception.
	* If 'ignore', then invalid parsing will return the input.
	.. note:: 'ignore' doesn't work yet when `arg` is pandas-on-Spark Series.

[ueshin]

Yes, eventually we should update the docs, but so far pandas 3 is not fully supported and I'm still not sure whether we can make it by 4.2.0 release, so I think we should keep it as-is for now.

Also I see another PR to show warning with pandas 3.

• [SPARK-55965][PYTHON] Add warning when pandas >= 3.0.0 is used with PySpark #54760

[dongjoon-hyun]

Got it~

[HyukjinKwon]

Looks good except one comment above.

[dongjoon-hyun]

+1, LGTM. Thank you, @ueshin .

[HyukjinKwon]

Merged to master.