[SPARK-54472][SQL] Add ORC read and write support for TIME type

[vinodkc]

What changes were proposed in this pull request?

This PR adds ORC serialization and deserialization support for Spark's TIME type.

Why are the changes needed?

TIME type currently lacks ORC support, preventing users from:

• Reading/writing ORC files with TIME columns
• Integrating TIME data with existing ORC-based data lakes
• Preserving TIME precision in columnar storage

Does this PR introduce any user-facing change?

Yes. Users can now:

1. Read ORC with TIME columns

spark.read.format("orc").load("data.orc")
// Returns DataFrame with TIME columns preserved

2. Write DataFrames with TIME to ORC

val df = spark.sql("SELECT TIME'14:30:45.123456' as shift_start")
df.write.format("orc").save("output.orc")

Technical Details

Storage Format

ORC Column:
  Physical Type: LONG (nanoseconds since midnight)
  Custom Attribute: spark.sql.catalyst.type = "time(<precision>)"
  Value Range: 0 to 86,399,999,999,999

Precision Handling

Precision	Catalyst Type	ORC Attribute	Example Value
0 (seconds)	TimeType(0)	"time(0)"	12:34:56
3 (millis)	TimeType(3)	"time(3)"	12:34:56.123
6 (micros)	TimeType(6)	"time(6)"	12:34:56.123456

Future Compatibility

• Versioned via file metadata: Uses existing org.apache.spark.version for compatibility
• Forward-compatible: If ORC adds native TIME type, can migrate based on version

How was this patch tested?

Added tests in OrcQuerySuite

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

Yes.
Generated-by: Claude 3.5 Sonnet

AI assistance was used for:

• Code pattern analysis and design discussions
• Implementation guidance following Spark conventions
• Test case generation and organization
• Documentation and examples

[dongjoon-hyun]

+1, LGTM. Thank you, @vinodkc .

[dongjoon-hyun]

Merged to master for Apache Spark 4.2.0.