#set default `exact` option

19 messages · Page 1 of 1 (latest)

wicked stream
#

Hey!
for most of my use-cases, I wanna do

refetchQueries({ queryKey: [ /* ... */ ], exact: true });

is there a way for me to set exact: true by default across the board? instead of setting separately?
couldn't find a way to do it by myself

lime oracle
#

nope. How do your queryKeys look like? A little restructure might help

wicked stream
#

well, generally the http call itself really inspires the key.
but generally speaking, most of the time I have a

const todosKeys = {
  all: [{ orgId }, 'todos'],
  single: (id: string) => ([...todosKeys.all, id])
}

The orgId is sort of a prefix for context,
while using my app, a user can be assocaited to several "organizations", and move between them (and they're completely unrelated with completely different data, that's why it felt important to contextualize).

So say, when I create a new todo
I wanna

  queryClient.setQueryData(single(newTodo.id), newTodo);
  queryClient.refetchQueries({ queryKey: all, exact: true });

or when I remove a todo, I only want to do

    onSettled: () => queryClient.refetchQueries({ queryKey: all, exact: true })

I don't want all todos in the world to be refetched

#

does this make sense? am I going with a "correct" approach?

#

Also - it doesn't really help that there're no examples / calrification as to what does exact: false match with 😰

lime oracle
#

So, if you refetch queryKey: todoKeys.all, I'd expect everything todo-related to be refetched (the list and the details). At least from reading the code. So I'd probably structure a bit more, like:

const todosKeys = {
  all: [{ orgId }, 'todos'],
  list: () => ([...todosKeys.all, 'list'])
  single: (id: string) => ([...todosKeys.all, 'single', id])
}

If I want to refetch the list, I'd do:

queryKey: todoKeys.list()

and then no "exact" is needed

wicked stream
#

but then every "single" todo would be refetched, no?

#

since that's what "inclusive" means, or am I misunderstanding?

if my filter is [{ orgId }, 'todos'], and say this is my cache (figuratively)

{
 [{ orgId }, 'todos']: ...all todos,
 [{ orgId }, 'todos', { id: 1 }]: ...todo1,
 [{ orgId }, 'todos', { id: 2 }]: ...todo2, 
}

all of these would match the given filter, right?

#

also - I didn't get why separating between all and list gives any benefit 😅
sorry for my cluelessness

lime oracle
#

none of them would match 🙃

#

let's take a step back

wicked stream
#

updated the filter 🥴

#

typo

lime oracle
#

and say this is your cache:

{
 [{ orgId }, 'todos', 'list']: ...all todos,
 [{ orgId }, 'todos', 'detail', { id: 1 }]: ...todo1,
 [{ orgId }, 'todos', 'detail', { id: 2 }]: ...todo2, 
}

you have one todo list, where all your todos are in, and two detail entries.

Now you can:

  • target all 3 with: [{ orgId }, 'todos'] (that's the ALL key)
  • target just the list with [{ orgId }, 'todos', 'list']
  • target both details with [{ orgId }, 'todos', 'detail']
  • target one detail by id with [{ orgId }, 'todos', 'detail', { id }]
wicked stream
#

ah ha,
so we in advance, notate our keys explicitly

#

I thought that "telling the story" of the data with just the params is good enough

#

interesting

lime oracle
#

it's about forming a hierarchy, depending on how you want to target them with query invalidation / refetches

wicked stream
#

I see