node.js - MongoDB aggregation limit matches to 1 of each

Question

I'm trying to use MongoDB aggregation get values that are in the idArray but doesn't exist as a doc:

 idArray = ["a", "b", "c"]

docs

     {node: "a", value: "3"}
     {node: "c", value: "3"}
     {node: "a", value: "1"}
     {node: "a", value: "4"}

   db.docs.aggregate([
      {$match:
        node: 
          $in: idArray}
      {$group: {"_id":null , "matches":{$addToSet:"$node"}} } 
      {$project:
        node_not_found: { $setDifference: [ INPUT_ARRAY, "$matches" ] }
        _id: 0}
    ]).each (err, node) ->

This should produce:

{ "node_not_found" : [ "b" ] }

My issue is that this is quite slow as it matches a multiple times and tries to push it all into the array. How do I limit the match to have only 1 of a so that duplicate work isn't done?

Answer 1

How do I limit the match to have only 1

I don't think there is better way to limit document in match stage using $in query.

My issue is that this is quite slow as it matches

You need to avoid such array operations from your query,

• remove last $project stage and do that operation in your client side,
• create single field index on node field, it will be more effective for performance,
• i am suggesting below query, i have not tested it but you can try may be it will help,

Suggestion 1:

  { $match: { node: { $in: ["a", "b", "c"] } } },
  { $group: { "_id": "$node" } }

Playground

Result:

[{ "_id": "c" }, { "_id": "a" }]

Suggestion 2:

  { $match: { node: { $in: ["a", "b", "c"] } } },
  { $group: { "_id": null, "matches": { $addToSet: "$node" } } }

Playground

Result:

[{ "_id": null, "matches": ["c", "a"] }]

Other stuff you can do in your client side code.

You can check state of your query using explain(), see the examples and you can try in your query like db.docs.explain().aggregate(..) it will provide each statistics of your query stages.