MultiplyYMInterval¶

Overview¶

The MultiplyYMInterval expression multiplies a year-month interval by a numeric value, returning a new year-month interval. This expression supports various numeric data types and handles precision conversions appropriately for each type.

Syntax¶

year_month_interval * numeric_value

Arguments¶

Argument	Type	Description
`interval`	YearMonthIntervalType	The year-month interval to be multiplied
`num`	NumericType	The numeric multiplier (byte, short, int, long, float, double, or decimal)

Return Type¶

YearMonthIntervalType() - Returns a year-month interval representing the multiplication result.

Supported Data Types¶

Interval: YearMonthIntervalType only
Numeric: All numeric types including:
Integer types: ByteType, ShortType, IntegerType, LongType
Floating-point types: FloatType, DoubleType
DecimalType (arbitrary precision)

Algorithm¶

Internally represents year-month intervals as integer months
Uses type-specific multiplication logic based on the numeric operand type
For integer types: Uses Math.multiplyExact() for overflow detection
For floating-point types: Uses DoubleMath.roundToInt() with HALF_UP rounding
For decimal types: Performs decimal multiplication then rounds to integer months
Applies exact conversion methods to prevent silent overflow/underflow

Partitioning Behavior¶

Preserves partitioning: Yes, this is a deterministic transformation that doesn't require data redistribution
Requires shuffle: No, the operation is performed locally on each partition

Edge Cases¶

Null handling: Returns null if either operand is null (nullIntolerant = true)
Overflow behavior:
Integer multiplication uses Math.multiplyExact() throwing ArithmeticException on overflow
Long results converted using Math.toIntExact() throwing exception if out of int range
Decimal results use intValueExact() throwing exception if fractional part exists after rounding
Floating-point precision: Uses HALF_UP rounding mode for consistent behavior
Zero multiplication: Results in zero-month interval (valid)

Code Generation¶

Supports full Tungsten code generation for all numeric types: - Integer types: Direct Math.multiplyExact() calls - Long type: Nested Math.toIntExact(Math.multiplyExact()) calls
- Floating-point types: DoubleMath.roundToInt() with rounding mode - Decimal types: Inline decimal arithmetic with rounding and exact conversion

Examples¶

-- Multiply interval by integer
SELECT INTERVAL '2-6' YEAR TO MONTH * 3;  -- Results in '7-6' (7 years 6 months)

-- Multiply interval by decimal
SELECT INTERVAL '1-0' YEAR TO MONTH * 2.5;  -- Results in '2-6' (2 years 6 months)

// DataFrame API usage
import org.apache.spark.sql.functions._

df.select(col("year_month_interval") * lit(2))
df.select(col("year_month_interval") * col("multiplier"))