BitwiseXor¶

Overview¶

BitwiseXor is a binary arithmetic expression that performs bitwise XOR operations between two integral expressions. It returns 1 for each bit position where the corresponding bits of either operand (but not both) are 1, and 0 where both bits are the same.

Syntax¶

expression1 ^ expression2

// DataFrame API
col("column1").bitwiseXOR(col("column2"))
// or using expr()
expr("column1 ^ column2")

Arguments¶

Argument	Type	Description
left	Expression	Left operand for the bitwise XOR operation
right	Expression	Right operand for the bitwise XOR operation

Return Type¶

Returns the same integral data type as the input expressions (after type coercion if needed).

Supported Data Types¶

ByteType
ShortType
IntegerType
LongType

All inputs must be of IntegralType (whole numbers only).

Algorithm¶

Type-specific XOR functions are lazily initialized based on the expression's data type
For each supported type, performs native Scala bitwise XOR operation (^)
Results are cast back to the appropriate type (.toByte, .toShort) for smaller integral types
Uses null-safe evaluation to handle null inputs
Implements commutative property optimization during query planning

Partitioning Behavior¶

Preserves partitioning: Yes, when used in projections
Requires shuffle: No, this is a row-level operation that doesn't require data movement
Can be pushed down to individual partitions for parallel execution

Edge Cases¶

Null handling: If either operand is null, the result is null (null-safe evaluation)
Empty input: Not applicable for binary operations
Overflow/underflow: Not applicable - bitwise operations don't overflow within their type bounds
Type coercion: Both operands must resolve to the same integral type after implicit casting

Code Generation¶

This expression extends BinaryArithmetic which supports Tungsten code generation. The expression will generate optimized Java bytecode rather than falling back to interpreted evaluation mode for better performance.

Examples¶

-- Basic bitwise XOR
SELECT 3 ^ 5;
-- Result: 6 (011 XOR 101 = 110)

-- Column operations  
SELECT col1 ^ col2 FROM table1;

-- With literals
SELECT id ^ 255 FROM users;

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

df.select(col("value1").bitwiseXOR(col("value2")))

// Using expression strings
df.selectExpr("value1 ^ value2 as xor_result")

// With literals
df.select(col("flags") ^ lit(128))