Tan¶

Overview¶

The Tan expression computes the trigonometric tangent of a numeric value. It is a unary mathematical expression that takes a single numeric input representing an angle in radians and returns its tangent value.

Syntax¶

TAN(expr)

// DataFrame API
import org.apache.spark.sql.functions.tan
df.select(tan($"angle_column"))

Arguments¶

Argument	Type	Description
child	Expression	A numeric expression representing an angle in radians

Return Type¶

Returns a DoubleType value representing the tangent of the input angle.

Supported Data Types¶

ByteType
ShortType
IntegerType
LongType
FloatType
DoubleType
DecimalType (converted to Double internally)

Algorithm¶

Inherits from UnaryMathExpression which provides the evaluation framework
Uses Java's math.tan function for the actual computation
Input values are converted to Double type before computation
Returns the tangent value as a Double result
Leverages Catalyst's expression evaluation infrastructure for optimization

Partitioning Behavior¶

This expression preserves partitioning characteristics:

Does not require shuffle operations
Maintains existing data partitioning since it's a row-level transformation
Can be pushed down in query optimization when appropriate

Edge Cases¶

Null handling: Returns null if the input expression evaluates to null
Infinity cases: Returns positive or negative infinity for angles where tangent is undefined (e.g., π/2, 3π/2)
NaN handling: Returns NaN (Not a Number) if input is NaN
Overflow behavior: Follows IEEE 754 floating-point standards for extreme values
Zero input: TAN(0) returns 0.0 as expected

Code Generation¶

This expression supports Catalyst code generation (Tungsten):

Inherits code generation capabilities from UnaryMathExpression
Generates efficient Java bytecode for the math.tan operation
Falls back to interpreted mode only in exceptional cases where code generation fails

Examples¶

-- Basic usage
SELECT TAN(0) AS tan_zero;
-- Result: 0.0

-- With column data
SELECT angle, TAN(angle) AS tangent_value
FROM trigonometry_table;

-- Using π/4 (45 degrees)
SELECT TAN(PI()/4) AS tan_45_degrees;
-- Result: 1.0

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

// Basic tangent calculation
df.select(tan(col("angle_radians")))

// With alias
df.select(tan(col("angle_radians")).alias("tangent_value"))

// Combined with other math functions  
df.select(
  col("angle"),
  tan(col("angle")).alias("tangent"),
  cos(col("angle")).alias("cosine")
)