what's the difference between ways to parameterize ngrx selector

Question

in ngrx, there are 2 ways(see following code) to parameterize a selector.

Question: what's the difference between them?

const getUser1 = () =>
  createSelector(
    getEntities,
    (state, props) => state[props.id]
  );

const getUser2 = id =>
  createSelector(
    getEntities,
    state => state[id]
  );

I think memoization works fine in both.

and article below says....

If the parameter doesn’t change over time we can use a factory function https://blog.angularindepth.com/ngrx-parameterized-selector-e3f610529f8

ok. these selectors are used as follows.

@Component({
  selector: 'app-user-container',
  template: ``
})
export class UserContainerComponent {
  user1$ = this.store.pipe(select(getUser1(), { id: 1 }))
  user2$ = this.store.pipe(select(getUser2(2)))

  constructor(private store: Store<AppState>) {}
}

in both case (getUser1() and getUser2(2)), factory functions are evaluated when intializing the class(so, parameter will not be changed for both selector). I couldn't find the difference in this case....

when is getUser1 useful?

Answer 1

Pre v6.2 it wasn't possible to use the first selector example, so the only way to pass props to a selector was to create a factory function.

Both examples in the question do exactly the same.

The difference would be that props would be passed down to all selectors in createSelector, the factory function would not.