[SPARK-52563][PS] Fix var naming bug in _assert_pandas_almost_equal

[petern48]

What changes were proposed in this pull request?

Small bug fix where the wrong variable names were used

Why are the changes needed?

The function uses lval and rval instead of the parameters val1 and val2

Does this PR introduce any user-facing change?

No

How was this patch tested?

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

No

[petern48]

First PR to spark. I would appreciate some help with creating the Jira issue. It didn't seem like I can create one myself or create a Jira account for Apache

[xinrong-meng]

Thanks for contribution!
Would you mind creating a JIRA ticket and linking the number to SPARK-XXXXX in your PR title? Please see https://spark.apache.org/contributing.html

[petern48]

Thanks for contribution! Would you mind creating a JIRA ticket and linking the number to SPARK-XXXXX in your PR title? Please see https://spark.apache.org/contributing.html

Yep, just did. I only recently realized I could request an account so I can create tickets.

[allisonwang-db]

Thanks for the fix! Would be great to add a test case for it.

[petern48]

@allisonwang-db I actually don't really see a good test we could write here? Do you? This function is actually a nested function (so more like a lambda) that is inside of a private function (_assert_pandas_almost_equal)... See here. To me, it doesn't make sense to write a test at this level.

Interestingly, this PR shouldn't change the behavior of the code at all since the variables lval and rval always happen to have the desired values of the intended arguments val1 and val2. This change is more of a typo fix to protect against errors if someone were to change the code later.

[allisonwang-db]

Looks like the inner function is used in this case:

for lval, rval in zip(left[lcol].dropna(), right[rcol].dropna()):
                if not compare_vals_approx(lval, rval):

[petern48]

Looks like the inner function is used in this case:

Yes, it certainly is used, but my point was that it's still nested inside the _assert_pandas_almost_equal() function, so I can't access it from a test without unnesting the function. Instead, I added some tests to cover other error types for _assert_pandas_almost_equal() that were previously missing.

@allisonwang-db What do you think?

[HyukjinKwon]

cc @xinrong-meng

[allisonwang-db]

@petern48 Thanks for fixing the bug and adding the test cases!

[HyukjinKwon]

Merged to master.