Summary

# GraphQL summary

You can now:

- Design schemas with nullability discipline and connections
- Author resolvers with per-request DataLoader to avoid N+1
- Build mutations with payload + UserError types
- Implement subscriptions over WebSocket or SSE
- Wire field-level authorisation in resolvers
- Cache on the client (Apollo / Relay) and server
- Operate persisted queries + depth limits in production
- Compose subgraphs with Apollo Federation v2

# Your next step - a per-request DataLoader

import DataLoader from 'dataloader';

function makeUserLoader(db) {
  return new DataLoader(async (ids) => {
    const rows = await db.users.findMany({ where: { id: { in: [...ids] } } });
    const byId = new Map(rows.map((r) => [r.id, r]));
    return ids.map((id) => byId.get(id) || null);
  });
}

// Create one PER REQUEST (never share across requests)
function createContext(req) {
  return { loaders: { user: makeUserLoader(db) } };
}

const resolvers = {
  Post: {
    author: (post, _, ctx) => ctx.loaders.user.load(post.authorId)
  }
};

# Next: federation, persisted queries, schema-first design at scale.