Sqrt¶

Overview¶

The Sqrt expression computes the square root of a numeric value. It wraps the standard math.sqrt function to provide square root functionality in Spark SQL expressions.

Syntax¶

SQRT(expr)

sqrt(col("column_name"))

Arguments¶

Argument	Type	Description
child	Expression	The numeric expression to compute the square root of

Return Type¶

Returns DoubleType - the square root result as a double-precision floating-point number.

Supported Data Types¶

Supports all numeric data types:

ByteType
ShortType
IntegerType
LongType
FloatType
DoubleType
DecimalType

Algorithm¶

The expression evaluation follows these steps:

Input value is converted to Double type if necessary
The standard math.sqrt function from Scala/Java Math library is applied
Result is returned as a Double value
Inherits null propagation behavior from UnaryMathExpression
Invalid inputs (negative numbers) return NaN according to IEEE 754 standard

Partitioning Behavior¶

This expression preserves partitioning characteristics:

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

Edge Cases¶

Null handling: Returns null when input is null (standard null propagation)
Negative numbers: Returns NaN (Not a Number) for negative input values following IEEE 754 specification
Zero input: Returns 0.0 for input value of 0
Positive infinity: Returns positive infinity for positive infinity input
NaN input: Returns NaN when input is already NaN

Code Generation¶

This expression supports Tungsten code generation for optimal performance. It extends UnaryMathExpression which provides code generation capabilities, avoiding interpreted mode execution in most cases.

Examples¶

-- Basic square root calculation
SELECT SQRT(4);
-- Result: 2.0

-- Square root of column values
SELECT name, SQRT(area) as side_length FROM squares;

-- Handling edge cases
SELECT SQRT(-1);  -- Returns NaN
SELECT SQRT(0);   -- Returns 0.0
SELECT SQRT(NULL); -- Returns NULL

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

df.select(sqrt($"number").as("square_root"))

// With column reference
df.withColumn("sqrt_value", sqrt(col("input_column")))