XORswap

XORswap, or the XOR swap trick, is a technique for swapping the values of two variables without using a temporary storage location. It relies on the bitwise exclusive OR operation to rearrange bits so that the original values are exchanged through a three-step sequence.

A common implementation uses three statements in this order: a = a ^ b; b = a ^ b; a = a ^ b;. After the first step, a holds the bitwise XOR of the original a and b. The second step assigns to b the XOR of this result with the new a, which yields the original a. The third step completes the swap by combining the new a with b to obtain the original b. This only works reliably for base-typed values that support XOR, typically integers.

But XORswap has notable drawbacks. If a and b refer to the same memory location (aliases), the method breaks and can yield zero. It also tends to be less readable and more error-prone than straightforward swapping with a temporary variable. In addition, it is not applicable to floating-point numbers or many non-integer types without reinterpretation of memory, which introduces further portability concerns. Modern processors and compilers are often able to optimize standard swap operations to use registers efficiently, making the XOR approach less advantageous in practice.

In contemporary programming, the standard approach—using a temporary variable or a language-provided swap construct (for example, a, b = b, a in Python or std::swap in C++)—is preferred for clarity and safety. The XOR swap technique remains primarily of historical interest and educational value, illustrating how bitwise operations can manipulate data.