[SPARK-51420][SQL] Get minutes of TIME datatype

[the-sakthi]

What changes were proposed in this pull request?

This PR adds support for extracting the minute component from TIME (TimeType) values in Spark SQL.

scala> spark.sql("SELECT minute(TIME'07:01:09.12312321231232');").show()
+------------------------------+
|minute(TIME '07:01:09.123123')|
+------------------------------+
|                             1|
+------------------------------+

Why are the changes needed?

• Spark previously supported minute() for only TIMESTAMP type values.
• TIME support was missing, leading to implicit casting attempt to TIMESTAMP, which was incorrect.
• This PR ensures that minute(TIME'HH:MM:SS.######') behaves correctly without unnecessary type coercion.

Does this PR introduce any user-facing change?

Yes

• Before this PR, calling minute(TIME'HH:MM:SS.######') resulted in a type mismatch error or an implicit cast attempt to TIMESTAMP, which was incorrect.
• With this PR, minute(TIME'HH:MM:SS.######') now works correctly for TIME values without implicit casting.
• Users can now extract the minute component from TIME values natively.

How was this patch tested?

By running new tests:

$ build/sbt "test:testOnly *TimeExpressionsSuite.scala"

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

No

[the-sakthi]

@MaxGekk looking forward to know your thoughts on this one!

[MaxGekk]

For sure, it would be nice to re-use the existing code for getting minutes, but I believe it would be better to implement minute from a TIME value as a separate expression.

[the-sakthi]

Thanks for the review @MaxGekk
That totally makes sense. I’ll introduce a separate MinuteForTime expression for TIME inputs that doesn’t depend on GetTimeField or time zones. For TIMESTAMP inputs, I’ll keep using the existing Minute expression (or rename it to MinuteForTimestamp). Then, in the function registry, I’ll dispatch to either MinuteForTime or MinuteForTimestamp based on the input data type. Let me know if that doesn’t align with what you had in mind.

[MaxGekk]

I’ll keep using the existing Minute expression (or rename it to MinuteForTimestamp).

@the-sakthi Let's leave it as is.

Then, in the function registry, I’ll dispatch to either MinuteForTime or MinuteForTimestamp based on the input data type.

Could you introduce new expression MinutesOfTime and place it to timeExpressions.scala (see #50287). MinutesOfTime should extend RuntimeReplaceable and use StaticInvoke. The last one just invokes getMinutesOfTime from DateTimeUtils.

[the-sakthi]

Let me know if this updated PR aligns better with your suggestions, @MaxGekk
Also, I'm looking for guidance, if possible, on fixing the codegen issue which I pointed above.

[MaxGekk]

@the-sakthi Could you update PR's description:

scala> spark.sql("SELECT minute(TIME'12:58:59');").show()
+-------------------+
|minute(46739000000)|

This PR #50299 fixed the column name.

[the-sakthi]

Thanks for the review @MaxGekk , I'll shortly in few mins update the PR. Major changes:

• Will remove the LiteralGenerator part of the change for a separate PR
• Will remove the checkConsistencyBetweenInterpretedAndCodegen part of the testing.
• Will cater to other review comments and update the PR description

[the-sakthi]

Updated the revision with the suggested changes @MaxGekk ! Let me know how this one looks.

[the-sakthi]

Moved the function usage/desc/example to the builder to fix a failing test which was complaining about this.

[the-sakthi]

Rebased with current main (master) branch!
Let me know how this revision looks, @MaxGekk !

[MaxGekk]

+1, LGTM. Merging to master.
Thank you, @the-sakthi.

[the-sakthi]

Thank you very much for helping on this one and merging @MaxGekk
Much appreciated!