RoundCeil¶

Overview¶

The RoundCeil expression performs ceiling-based rounding on a numeric value to a specified number of decimal places. It rounds the input value away from zero to the nearest value at the given scale using BigDecimal.RoundingMode.CEILING.

Syntax¶

ROUND_CEILING(value, scale)

// DataFrame API usage would typically go through SQL functions

Arguments¶

Argument	Type	Description
child	Expression	The numeric value to be rounded
scale	Expression	The number of decimal places to round to (must be integer type)

Return Type¶

Returns a decimal type that preserves the precision and scale characteristics of the input decimal value after applying ceiling rounding.

Supported Data Types¶

Input value (child): DecimalType only
Scale parameter: IntegerType only

Algorithm¶

Evaluates the child expression to get the decimal value to be rounded
Evaluates the scale expression to determine the number of decimal places
Applies BigDecimal.RoundingMode.CEILING to round the value away from zero
Returns the rounded decimal result
Inherits core rounding logic from the RoundBase parent class

Partitioning Behavior¶

This expression preserves partitioning characteristics:

Does not require data shuffle as it operates on individual row values
Maintains existing partitioning scheme since it's a deterministic row-level transformation
Can be pushed down to individual partitions for parallel execution

Edge Cases¶

Null handling: If either the child value or scale parameter is null, the expression returns null
Scale bounds: The scale parameter must be a valid integer; behavior with extreme scale values depends on underlying BigDecimal implementation
Precision limits: Result precision is constrained by Spark's DecimalType precision limits
Ceiling rounding: Always rounds away from zero (e.g., 2.1 becomes 3.0, -2.1 becomes -2.0)

Code Generation¶

This expression extends RoundBase which likely supports Catalyst code generation for optimized execution, though the specific code generation behavior depends on the parent class implementation.

Examples¶

-- Round 3.14159 to 2 decimal places using ceiling
SELECT ROUND_CEILING(3.14159, 2); -- Returns 3.15

-- Round to whole number
SELECT ROUND_CEILING(2.1, 0); -- Returns 3

-- Negative number ceiling rounding  
SELECT ROUND_CEILING(-2.7, 0); -- Returns -2

// DataFrame API usage would typically use SQL expression
import org.apache.spark.sql.catalyst.expressions.RoundCeil
import org.apache.spark.sql.catalyst.expressions.Literal

// Direct expression usage in Catalyst
val roundExpr = RoundCeil(
  child = Literal(3.14159),
  scale = Literal(2)
)