Implement GROUP BY about join-monster HOT 3 OPEN

Hi,

I am developing a prototype for my current project and man, join-monster is a decent piece of software that can solve 80% of my problems out of the box. GroupBy is in the other 20%, so I was wondering why this feature was closed?

I hacked a group by into the QueryRoot resolver by just collecting the fields to be grouped by in the context and appending groupby in the dbCall function of the demo - it works pretty well when you sqlBatch the join. So I was wondering if there's any blocker further down the road that I've missed? If not, a hint into the right direction how this ought to be tackled would be highly appreciated. I'd be happy to contribute to this great project :)

Thanks!

from join-monster.

Interesting.... do you have a sample or a fork I can check out? I ended up closing it because I couldn't get a consistent, efficient implementation that worked in all situations, e.g. after a sqlJoin or a sqlBatch. I am certainly open to the idea.

from join-monster.

I couldn't get a consistent, efficient implementation that worked in all situations

I agree, adding robust support without limitations will be hard, if possible at all. Please do not consider this a proper approach - it's just the result of a quick PoC if join-monster can solve problems in my current project - I added a groupBy hack to the demo here, just

npm install
npm start

Sample query:

{
  users {
    id
    stats(from: "2016-01-01", to: "2017-08-02") {
      comments
      likes
    }
    posts {
      body
      stats(from: "2016-01-01", to: "2017-08-02") {
        comments
        likes
      }
    }
  }
}

General functionality is to populate context.groupBy in the where() of the StatsConnector by climbing up the nodes in the sqlASTNode and then appending the group by SQL in the resolver() of the QueryRoot.

from join-monster.