relayjs - Relationship between GraphQL and database when using connection and pagination? -

it easy set pagination relay there's small detail unclear me.

both of relevant parts in code marked comments, other code additional context.

const posttype = new graphqlobjecttype({   name: 'post',   fields: () => ({       id: globalidfield('post'),       title: {         type: graphqlstring       },   }),   interfaces: [nodeinterface], })  const usertype = new graphqlobjecttype({   name: 'user',   fields: () => ({       id: globalidfield('user'),       email: {         type: graphqlstring       },       posts: {         type: postconnection,         args: connectionargs,         resolve: async (user, args) => {           // getuserposts() in next code block -> gets data db           // pass args (e.g "first", "after" etc) , user id (to user posts)           const posts = await getuserposts(args, user._id)           return connectionfromarray(posts, args)         }       },   }),   interfaces: [nodeinterface], })  const {connectiontype: postconnection} =                connectiondefinitions({name: 'post', nodetype: posttype}) 

---

exports.getuserposts = async (args, userid) => {     try {       // using mongodb , mongoose question relevant every db       // .limit() -> how many posts return       const posts = await post.find({author: userid}).limit(args.first).exec()       return posts     } catch (err) {       return err     } } 

---

cause of confusion:

• if pass first argument , use in db query limit returned results, hasnextpage always false. efficient breaks hasnextpage (haspreviouspage if use last)
• if don't pass first argument , don't use in db query limit returned results, hasnextpage working expected return all items queried (could thousands)
  • even if database on same machine (which isn't case bigger apps), seems very, very, inefficient , awful. please prove me im wrong!
  • as far know, graphql doesn't have server-side caching therefore there wouldn't point return results (even if did, users don't browse 100% content)

---

what's logic here?

one solution comes mind add +1 first value in getuserposts, retrieve 1 excess item , hasnextpage work. feels hack , there's excess item returned - grow relatively if there many connections , requests.

are expected hack that? expected return all results?

or did misunderstand whole relationship between database , grahpql / relay?

---

what if used fb dataloader , redis? change logic?

cause of confusion

the utility function connectionfromarray of graphql-relay-js library not solution kinds of pagination needs. need adapt our approach based on our preferred pagination models.

connectionfromarray function derives values of hasnextpage , hasprevisouspage given array. so, observed , mentioned in "cause of confusion" expected behavior.

as confusion whether load data or not, depends on problem @ hand. loading items may make sense in several situations such as:

• the number of items small , can afford memory required store items.
• the items requested , need cache them faster access.

two common pagination models numbered pages , infinite scrolling. graphql connection specification not opinionated pagination model , allows both of them.

for numbered pages, can use field totalpost in graphql type, can used display links numbered pages on ui. on back-end, can use feature skip fetch needed items. field totalpost , current page number eliminates dependency on hasnextpage or haspreviouspage.

for infinite scrolling, can use cursor field, can used value after in query. on back-end, can use value of cursor retrieve next items (value of first). see example of using cursor in relay documention on graphql connection. see this answer graphql connection , cursor. see this , this blog posts, better understand idea of cursor.

---

what's logic here?

are expected hack that?

no, ideally we're not expected hack , forget it. leave technical debt in project, cause more problems in long term. may consider implementing own function return connection object. ideas of how in implementation of array-connection in graphql-relay-js.

is expected return results?

again, depends on problem.

---

what if used fb dataloader , redis? change logic?

you can use facebook dataloader library cache , batch-process queries. redis option caching results. if load (1) items using dataloader or store items in redis , (2) items lightweight, can create array of items (following kiss principle). if items heavy-weight, creating array may expensive operation.