[SPARK-55293][PS][TESTS] Avoid using old offset aliases

[ueshin]

What changes were proposed in this pull request?

Avoid using old offset aliases.

• "H" -> "h"
• "T" -> "min"
• "S" -> "s"
• "L" -> "ms"
• "U" -> "us"
• "Q" -> "QE"
• "Y" -> "YE"

https://pandas.pydata.org/docs/user_guide/timeseries.html#offset-aliases

Why are the changes needed?

#54018 fixed the offset alias M to ME.

There are other aliases that are deprecated in pandas 2.2 / removed in pandas 3.0.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Fixed the related tests.

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

No.

[github-actions]

JIRA Issue Information

=== Sub-task SPARK-55293 ===
Summary: Avoid using old offset aliases
Assignee: None
Status: Open
Affected: ["4.2.0"]

---

This comment was automatically generated by GitHub Actions

[ueshin]

cc @gaogaotiantian @HyukjinKwon @zhengruifeng

[ueshin]

? My repo doesn't link the build status back.

https://github.com/ueshin/apache-spark/actions/runs/21533781589

[gaogaotiantian]

I believe there is a difference between an offset alias and a period alias. For example, period_range should use period_alias which should be M instead of ME. I'm not 100% sure which one is which in our code base.

I think the difference is that, offset alias means the next point after this much time; whereas period alias means the span of this much time.

[ueshin]

@gaogaotiantian Good catch!
I missed it because actually the test is skipped as it doesn't support Period type.

[HyukjinKwon]

Merged to master.