[WIP][SPARK-29669][SQL] Refactor IntervalUtils.fromDayTimeString()

[MaxGekk]

What changes were proposed in this pull request?

In the PR, I propose to refactor code related to the IntervalUtils.fromDayTimeString() method by:

• making it more generic and independent from the requested range of interval units
• introducing enumeration for unit names. This should allow to avoid typos in unit names.
• extracting unit value properties from fromDayTimeString() like min/max values and a function for converting parsed values to microseconds using exact math methods.

This PR includes code from #26313

Why are the changes needed?

• To improve maintainability of fromDayTimeString() and related code in IntervalUtils
• Avoid adhoc code

Does this PR introduce any user-facing change?

No

How was this patch tested?

By existing test in IntervalUtilsSuite and DDLParserSuite.

[MaxGekk]

This code has been borrowed from #26313

[MaxGekk]

The changes from #26313

[MaxGekk]

@dongjoon-hyun @srowen @cloud-fan @HyukjinKwon Could you take a look at this, please.

[SparkQA]

Test build #112960 has finished for PR 26327 at commit c3fe563.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[MaxGekk]

The build #26327 (comment) fails on: '1 2:03' day to hour because groups in the regexp are floating:

1. hour can be bound to group 5

   spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/IntervalUtils.scala

   Line 210 in dc987f0

   hours = toLongWithRange("hour", m.group(5), 0, 23)

2. or to group 6:

   spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/IntervalUtils.scala

   Line 217 in dc987f0

   hours = toLongWithRange("hour", m.group(6), 0, 23)

   
   The is a weird regexp :-E

[SparkQA]

Test build #113044 has finished for PR 26327 at commit f381103.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
• trait HasRelativeError extends Params
• class _ImputerParams(HasInputCol, HasInputCols, HasOutputCol, HasOutputCols, HasRelativeError):
• class _RobustScalerParams(HasInputCol, HasOutputCol, HasRelativeError):
• class HasRelativeError(Params):
• case class ShowColumnsStatement(
• case class DataSourceV2ScanRelation(
• case class OptimizeLocalShuffleReader(conf: SQLConf) extends Rule[SparkPlan]

[SparkQA]

Test build #113052 has finished for PR 26327 at commit 140a26f.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.