Hacker News Reader

Comment by nsxwolf | Hacker News Reader

I want to say something nice about this... but unless your entire enterprise is a pure CRUD app with just a SQL database I don't see how this is useful. Complex application backends pull data from multiple sources - SQL and NoSQL databases, REST and SOAP endpoints, message queues, service buses - and perform all sorts of transactions and data transformations.

One of the strengths of GraphQL is you can transform data from multiple sources into a common model the UI can use. If your whole backend is a single SQL database schema GraphQL doesn't seem to add a whole lot of value there. I'd rather have a code generator that builds a simple Go microservice and use protobuf JS in my front end.

View on HN

> I'd rather have a code generator that builds a simple Go microservice and use protobuf JS in my front end.

You might, but there's something to be said for consistency within an org. GraphQL is powerful and there's many a use case for querying data via a simple GraphQL API in the frontend. You could have just stated you have a bias against use of GraphQL without the need for multiple data sources, and left out the rest of the FUD.

Projects like Postgraphile, Hasura, AppSync, and Supabase's pg-graphql aren't popular because people are lazy or are out there in droves implementing APIs laden with antipatterns. They're popular because they're useful and powerful. And for many companies that funnel all of their data into single stores like Postgres or other relational databases, these kinds of tools are invaluable.

You can do this with GraphJin. We support adding custom resolvers currently we have an HTTP resolver built in which allows you to join your DB GraphQL query with a remote HTTP Rest endpoint. For example you want users from your db and their subscription data from Stripe. It works out of the box, GraphJin will write out the correct Rest call using data from the db call eg stripe_customer_id from the db will be inserted into the rest calls. Additionally its smart enough to handle rate limits and parallel fetches if required.

Here's an example https://pkg.go.dev/github.com/dosco/graphjin/core#example-pa...

And here's an example of a Redis custom resolver to join with data from Redis. https://pkg.go.dev/github.com/dosco/graphjin/core#Resolver

Reply on news.ycombinator.com