Usage of Prime() about dataloader HOT 7 CLOSED

After thinking about this a bit... It seems like priming should take place in the resolvers? Or maybe we are missing some kind of promise functionality on Load()... Curious to get your thoughts @nicksrandall

from dataloader.

A good example of using prime is here... https://github.com/facebook/dataloader#loading-by-alternative-keys

In that case, they are using prime method in the batch function to "prime" the cache by an alternate id.

from dataloader.

That seems like a reasonable way to do it. How would you recommend I get the instance of the dataloader to supply to my batch function (given that you need to supply a batch function to get an instance of a dataloader)?

from dataloader.

Off the top of my head I would declare variables of type dataloader.Interface that you can reference in your batch func and then just make sure that those variables are initialized before they are actually used.

from dataloader.

Here's some rough code... is this close to what you are suggesting?

type MyLoader struct {
  client myservice.Client
}

func (l *MyLoader) LoadByIDs(ctx context.Context, dl *dataloader.Interface) dataloader.BatchFunc {
  return func(ids []string) []*dataloader.Result {
    // fetch resources by ids, implementation omitted ... 
    if dl != nil {
      dl.Prime(resource.Name, resource)
    }
    // .... omit rest of implementation 
  }
}

// When handling requests, attach data loaders to the request context.
loader := MyLoader{client}
var dl *dataloader.Loader

dl = dataloader.NewBatchedLoader(loader.Load(ctx, dl))
ctx = context.WithValue(ctx, key, dl)

from dataloader.

@tonyghita yeah, something like that should work. I personally keep all my loaders on my context but essentially I do the same thing.

from dataloader.

Closing this.. feel free to reopen if you still have questions.

from dataloader.