Is it okay to just use JPA entities as a "shell" to hold primary keys or is that a bad practice?

Question

For example, lets say I want to assign a player to a team using the user and team's primary keys:

myEjb.java

public void assignPlayer(Integer teamId, Integer playerId) {
    Team team = em.find(Team.class, teamId);
    team.getPlayers().add(new Player(playerId));
    em.merge(team);
    .
    .
}

Is it wrong to do this? I feel like I'm saving myself an extra DB read by wrapping the player primary key in the object instead of first doing:

Player player = em.find(Player.class, playerId)

Answer 1

No, it's not a good idea.

First, the call to merge() is not necessary.

Second, there is a standard method to create an entity from an ID, without executing any query: getReference().

Team team = em.find(Team.class, teamId);
team.getPlayers().add(em.getReference(Player.class, playerId));

At least with getReference(), you get a real, managed entity. And if the code, later in the transaction, happens to call a method on that player, it won't get nullas a result: JPA will load the player state, and return the correct value.