DataLoader is a generic utility to be used as part of your application's data fetching layer to provide a consistent API over various backends and reduce requests to those backends via batching and caching.
On the back-end, it makes for nice documentation, but my sense is that things get pretty complicated when you need to support a graphQL query that may include multiple parents/descendants. Keeping your database queries clean means using a dataloader to avoid n+1 issues.
Apollo GraphQL: Should I include Plaid's API within GraphQL or separate into individual HTTP endpoints?
1 project | reddit.com/r/graphql | 23 Feb 2022
Caching data with DataLoader
1 project | dev.to | 26 Jan 2022
C++ and GraphQL
1 project | reddit.com/r/graphql | 17 Jan 2022
Graph QL query involving more content
2 projects | reddit.com/r/graphql | 8 Sep 2021
Do resolvers cause a lot of extra database transactions?
2 projects | reddit.com/r/graphql | 30 Jun 2021