There are strong recommendations to flatten out the state tree for an ngrx store, and I've worked to do so throughout my app. I understand with, and agree with the reasons behind doing so- flat state trees are easier to manage, deep copy, etc.
However, I've run into a scenario that leaves me wondering the most maintainable way of setting up my state tree. I have a feature, Crews which contains a number of users on a Crew. An instance of a crew can have substitute users (e.g.- if someone on that crew is unavailable). A sub simply defines the original user ID, and the replacement user ID. As far as I can tell, there are two ways I can do this. The most straightforward way would be to have a non-flat structure- something that might look like this:
export interface User {
id: string;
username: string;
}
export interface Crew {
id: string;
members: string[]; //User IDs
subs: [string, string][]; //[Orig UID, Sub UID]
}
export interface UserState extends EntityState<User> { }
export interface CrewState extends EntityState<Crew> { }
export interface AppState {
users: UserState;
crews: CrewState;
}
The other option, is to create a Subs feature, and the subs within Crew would point to an id in subs:
export interface User {
id: string;
username: string;
}
export interface Crew {
id: string;
members: string[]; //User IDs
subs: string[]; //Sub ID
}
export interface Sub {
id: string;
original: string; //Original UID
sub: string; //Sub UID
}
// Using ngrx entities
export interface UserState extends EntityState<User> { }
export interface CrewState extends EntityState<Crew> { }
export interface SubState extends EntityState<Sub> { }
export interface AppState {
users: UserState;
subs: SubState;
crews: CrewState;
}
But this seems a bit overkill to me- creating an entire new feature, simply to represent two string ID values. Any thought on which direction would be most maintainable/straightforward, or a better way to represent this?
