Asinh¶

Overview¶

The Asinh expression computes the inverse hyperbolic sine (arc hyperbolic sine) of a numeric value. It returns the hyperbolic angle whose hyperbolic sine is the input value, handling special cases like negative infinity appropriately.

Syntax¶

ASINH(expr)

// DataFrame API
col("column_name").asinh()
// or using expr
expr("asinh(column_name)")

Arguments¶

Argument	Type	Description
child	Expression	The numeric expression for which to calculate the inverse hyperbolic sine

Return Type¶

Double - returns a double-precision floating-point number representing the inverse hyperbolic sine.

Supported Data Types¶

Numeric data types that can be converted to Double:

TINYINT
SMALLINT
INTEGER
BIGINT
FLOAT
DOUBLE
DECIMAL

Algorithm¶

Uses the mathematical formula: asinh(x) = log(x + sqrt(x² + 1))
Implements special case handling for negative infinity input
Leverages StrictMath.log() for precise logarithmic calculations
Combines standard math operations (sqrt, addition) for the core computation
Maintains IEEE 754 compliance for floating-point edge cases

Partitioning Behavior¶

This expression preserves partitioning:

Does not require data shuffle as it's a unary transformation
Can be applied per-partition independently
Maintains the same number of rows and partitioning scheme

Edge Cases¶

Null handling: Returns null if input is null (inherited from UnaryMathExpression)
Negative infinity: Explicitly handled to return Double.NegativeInfinity
Positive infinity: Returns Double.PositiveInfinity through natural mathematical evaluation
NaN input: Returns NaN following IEEE 754 standards
Zero input: Returns 0.0 as asinh(0) = 0
Very large values: May approach positive/negative infinity asymptotically

Code Generation¶

This expression supports Tungsten code generation (Whole Stage Code Generation):

Implements doGenCode() method for compiled code path
Generates optimized Java bytecode avoiding function call overhead
Falls back to interpreted mode only when code generation is disabled
Uses direct Java math operations in generated code for maximum performance

Examples¶

-- Basic usage
SELECT ASINH(1.0);
-- Returns: 0.8813735870195430

-- Special case: zero
SELECT ASINH(0);
-- Returns: 0.0

-- Negative values
SELECT ASINH(-2.0);
-- Returns: -1.4436354751788103

-- With column data
SELECT name, ASINH(value) as asinh_value FROM measurements;

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

// Apply to column
df.select(col("value"), asinh(col("value")).alias("asinh_value"))

// Using expr function
df.selectExpr("value", "asinh(value) as asinh_value")

// In transformations
df.withColumn("asinh_transformed", asinh(col("numeric_column")))