I'm building a cache implementation in Java using Redisson. I want to use it to cache a numerical value. So I'm using getAtomicLong() like so:
RAtomicLong userNumber = redissonClient.getAtomicLong("my-key");
long value = userNumber.get();
However, the docs aren't very descriptive about what happens here, and so I have a few questions:
I wrote a silly Java example program and did some experimenting, so let me answer both the question you actually asked and then question I think you were trying to ask.
Assume that "my-key" does not yet exist in the cache. What does getAtomicLong() return?
An instance of RAtomicLong- because you might do something with the key/value in the future (like incrementAndGet or something).
If "my-key" does not exist, what does userNumber.get() return?
Zero. Yes, not null, yes not an exception, just 0. This was reasonably surprising in my test program.
The real interesting part about the Reddison API is that it leans hard on the atomic stuff - great if you could be updating a value from multiple threads - but seems to not document the simpliest use case: I want to read or write from Redis and it's not a number / not atomic / my threads or data is structured in such a way that they won't clobber each other.
That seems to be what Reddison's RBucket stuff is for.
That will return a null if the object is not in Redis.
RBucket<String> back = client.getBucket("foo");
String value = back.get();
if (value == null) {
System.out.println("NOPE, NULL");
} else {
System.out.println(value);
}
I really wish that would have been documented better - looking back I see "bucket" as a container for stuff, but for a while I assumed it meant some advanced Redis pattern and not "a holder for a value of a generic type".
(If you really do want the fancy stuff, there's an excellent baeldung artigle on it ).
