rethrow

Rethrow is the act of throwing an exception again after it has been caught in an error-handling routine. The primary purpose is to propagate the error to higher levels of the call stack or to defer handling to a different layer while preserving the original context. Rethrowing is distinct from wrapping or chaining, where a new exception is created and the original is attached as a cause; rethrow typically reuses the same exception object.

Rethrowing is commonly used after performing cleanup or after adding local context that does not resolve the error. It helps maintain visibility of the original fault, avoids swallowing the error, and allows higher-level code to apply its own handling policy.

Different languages treat rethrow and stack traces differently. In Java, rethrowing the caught exception with throw e preserves the original stack trace. Python’s except-block semantics allow rethrow via raise to keep the original traceback, whereas raise e can replace it. In C#, the preferred form to preserve the traceback is throw; whereas throw ex; resets the trace to the point of the throw.

Common considerations: avoid suppressing exceptions; ensure that any added context is informative; consider whether rethrowing plus additional handling (like logging) is appropriate, or if the error should be wrapped to convey higher-level meaning.

See also: exception handling, exception propagation, exception chaining, error handling best practices.