See: Description
| Interface | Description |
|---|---|
| Result<S,F> |
Encapsulates a single non-null value, representing the outcome of an operation that can either succeed or
fail.
|
Wave goodbye to slow exceptions and embrace clean, efficient error handling by encapsulating operations that may succeed or fail in a type-safe way.
Result objects represent the outcome of an operation, removing the need to
check for null. Operations that succeed produce results encapsulating a success value; operations
that fail produce results with a failure value. Success and failure can be represented by whatever types
make the most sense for each operation.
In Java, methods that can fail typically do so by throwing exceptions. Then, exception-throwing methods are called
from inside a try block to handle errors in a separate catch block.
int getServerUptime() {
int hours;
try {
Server server = connect();
server.refresh();
hours = server.getUptime();
} catch (ConnectionException exception) {
logger.error(exception.getMessage());
hours = -1;
}
return hours;
}
This approach is lengthy, and that's not the only problem — it's also very slow.
Conventional wisdom says exceptional logic shouldn't be used for normal program flow. Results make us deal with expected error situations explicitly to enforce good practices and make our programs run faster.
Let's now look at how the above code could be refactored if connect() returned a
Result object instead of throwing an exception.
int getServerUptime() {
final Result<Server, String> result = connect();
result.ifSuccess(Server::refresh);
result.ifFailure(logger::error);
return result.mapSuccess(Server::getUptime).orElse(-1);
}
In the example above, we used only 4 lines of code to replace the 10 that worked for the first one. But we can
effortlessly make it shorter by chaining methods. In fact, since we were returning -1 just to signal that the
underlying operation failed, we are better off returning a Result object
upstream. This will allow us to compose operations on top of getServerUptime() just like we did with connect().
int getServerUptime() {
return connect().ifSuccessOrElse(Server::refresh, logger::error)
.mapSuccess(Server::getUptime);
}
Result objects are immutable, providing thread safety without the need for
synchronization. This makes them ideal for multi-threaded applications, ensuring predictability and eliminating side
effects.
Result