I’m looking at a couple of designs for DAO interfaces. One has a single update() method, while the other has separate create() and update() methods.
Assuming that it’s possible to do something along the lines of ‘insert otherwise update’, what’s the benefit of a separate create() method? Is there something to do with concurrency lurking there?
Thanks in advance for your help.
Cheers,
Ross
I use two seperate methods when creating DAO’s simply for Clarity and Explicitness.
If there is a single Update method and somebody passes in an existing object, it will be updated. But what if the person really wanted to Create a new object and the fact that it already existed is an error?
With a single Update method, there’s really no way for anybody to tell what EXACTLY is going on behind the scenes.