[SPARK-52617][SQL]Cast TIME to/from TIMESTAMP_NTZ

[SubhamSinghal]

What changes were proposed in this pull request?

This PR adds support for casting TIME to/from TIMESTAMP_NTZ type

Why are the changes needed?

Adds casting support between 2 types.

Does this PR introduce any user-facing change?

yes

How was this patch tested?

By running the related test suites: CastSuiteBase

By manual tests:

timestamp_ntz to time conversion:

scala> spark.sql(
     |   """
     |     |SELECT CAST(CAST('2025-07-05 12:34:56.123456' AS TIMESTAMP_NTZ) AS TIME(6))
     |   """.stripMargin
     | ).show(false)

+------------------------------------------------------------------+
|CAST(CAST(2025-07-05 12:34:56.123456 AS TIMESTAMP_NTZ) AS TIME(6))|
+------------------------------------------------------------------+
|12:34:56.123456                                                   |
+------------------------------------------------------------------+

time to timestamp_ntz conversion:

scala> spark.sql("""
     |   SELECT CAST(CAST('12:34:56.123456' AS TIME(6)) AS TIMESTAMP_NTZ)
     | """).show(false)
+-------------------------------------------------------+
|CAST(CAST(12:34:56.123456 AS TIME(6)) AS TIMESTAMP_NTZ)|
+-------------------------------------------------------+
|2025-07-05 12:34:56.123456                             |
+-------------------------------------------------------+

logical plan analysis:

scala> spark.sql("SELECT CAST(make_time(12, 30, 40) AS TIMESTAMP_NTZ)").explain(true)

== Analyzed Logical Plan ==
make_timestamp_ntz(current_date(), make_time(12, 30, 40)): timestamp_ntz
Project [make_timestamp_ntz(current_date(Some(Asia/Kolkata)), make_time(12, 30, cast(40 as decimal(16,6)))) AS make_timestamp_ntz(current_date(), make_time(12, 30, 40))#0]
+- OneRowRelation

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

yes
Generated / corrected with AI assistance: chatGPT

[mihailotim-db]

On a separate note, is the new rule necessary? Should we try and place this in existing type coercion rules?

[subham611]

On a separate note, is the new rule necessary? Should we try and place this in existing type coercion rules?

@mihailotim-db Should I make RewriteTimeCastToTimestampNTZ extend TypeCoercionRule instead of Rule[LogicalPlan] and place RewriteTimeCastToTimestampNTZ inside CombinedTypeCoercionRule?