`persist()` makes a transient entity managed — it's meant for genuinely new entities and throws an exception if called on an entity that already has an assigned identifier matching an existing row (in most implementations). `merge()` takes a detached entity (or one with a set ID) and returns a new, managed copy with its state copied over — it doesn't attach the object you passed in itself, but rather a managed proxy/copy of it, which is why the returned reference from merge() must be used going forward, not the original object.

Advanced7 min

AmazonGoogleMicrosoftFlipkartOracleAdobe

#persist()#merge()

What is the difference between persist() and merge() in JPA?

Calling merge() on the Wrong Reference Is a Very Common Mistake

Because merge() returns a separate managed instance rather than making the passed-in object itself managed, code that continues modifying the original `detachedUser` reference after calling merge() is silently modifying an object the persistence context no longer cares about — the fix is always to reassign and continue working with merge()'s return value, not the original detached object.

merge() Returns a Different Reference

java

User detachedUser = getDetachedUserSomehow();
detachedUser.setName("New Name");
User managedUser = entityManager.merge(detachedUser); // managedUser != detachedUser
// changes must be made through managedUser going forward to be tracked