MongoDB $elemMatch projection on Nested Arrays

Question

I have a collection (summary) like this.

    {
    "id":"summaryid",
    "locations": [
        {
            "id": "loc1",
            "datacenters": [
                {
                    "id": "dc1.1",
                    "clusters": [
                        {
                            "id": "cl1.1",
                            "servers": [
                                {
                                    "id": "srvr1.1",
                                    "services": [
                                        {
                                            "id": "srvc1.1"
                                        }
                                    ]
                                }
                            ]
                        }
                    ]
                },
                {
                    "id": "dc1.2",
                    "clusters": [
                        {
                            "id": "cl1.2",
                            "servers": [
                                {
                                    "id": "srvr1.2",
                                    "services": [
                                        {
                                            "id": "srvc1.2"
                                        }
                                    ]
                                }
                            ]
                        }
                    ]
                }
            ]
        },
        {
            "id": "loc2",
            "datacenters": [
                {
                    "id": "dc2.1",
                    "clusters": [
                        {
                            "id": "cl2.1",
                            "servers": [
                                {
                                    "id": "srvr2.1",
                                    "services": [
                                        {
                                            "id": "srvc2.1"
                                        }
                                    ]
                                }
                            ]
                        }
                    ]
                },
                {
                    "id": "dc2.2",
                    "clusters": [
                        {
                            "id": "cl2.2",
                            "servers": [
                                {
                                    "id": "srvr2.2",
                                    "services": [
                                        {
                                            "id": "srvc2.2"
                                        }
                                    ]
                                }
                            ]
                        }
                    ]
                }

            ]
        }
    ]
}

Now I want only the clusters that are for datacenter with id dc1.1. I would like to exclude servers for the clusters.

I have tried using find query with $elemMatch and projections as below.

db.summary.find({}, {"locations": { $elemMatch: { "datacenters._id" : 
"dc1.1" } }, "locations.datacenters.clusters":0, 
"locations.datacenters.servers":0, "locations.datacentercount" : 0, 
"locations.clustercount" : 0, "locations.servercount" : 0}).pretty()

I am still getting all the datacenters instead of just 1 that matches the id. I am not sure if I am doing this right.

Thank you!

Can you post your expected json result? are you hoping for just { "id": "dc1.1", "clusters": [ {"id": "cl1.1",} ] } — bgraham
– bgraham, Commented Oct 10, 2017 at 15:49
Your $elemMatch instruction is using "datacenters._id" but the correct path (based on the example document you supplied) is "datacenters.id" — glytching
– glytching, Commented Oct 10, 2017 at 16:06
find is to find documents, not parts of the documents. Projection applies to all properties unconditionally, and is not the right tool here. You need to use aggregation to retrieve sub-documents. — Alex Blex
– Alex Blex, Commented Oct 10, 2017 at 16:13
Thank you Alex. I tried the aggregations and was get what I want. Following is what I am using. ` db.summary.aggregate([ {"$unwind": "$locations"}, {"$unwind": "$locations.datacenters"}, { $match: { "locations.datacenters._id" : "dc1.1" } }, { $project : {"locations.datacenters.servers" : 0, "locations.datacenters.clusters.servers" : 0} } ]).pretty() ` — Ed goo
– Ed goo, Commented Oct 10, 2017 at 16:42
New challenge is that I am using spring-mongodb-data's ProjectionOperation along with mongoTemplate. I keep getting the following error. ProjectionOperation projectStage = new ProjectionOperation().andExclude("locations.datacenters.servers", "locations.datacenters.clusters.servers"); Exclusion of field locations.datacenters.servers not allowed. Projections by the mongodb aggregation framework only support the exclusion of the _id field! — Ed goo
– Ed goo, Commented Oct 10, 2017 at 16:46

s7vr · Accepted Answer · 2017-10-12 13:18:23Z

6

It is not possible with $elemMatch to project the nested array element.

You can try the below aggregation in 3.4 server.

Use $unwind couple of times to reach the nested array and apply $match to pick the nested array element.

   db.summary.aggregate([
      {
        "$match": {
          "locations.datacenters._id": "dc1.1"
        }
      },
      {
        "$unwind": "$locations"
      },
      {
        "$unwind": "$locations.datacenters"
      },
      {
        "$match": {
          "locations.datacenters._id": "dc1.1"
        }
      },
      {
        "$project": {
          "locations.datacenters.clusters.servers": 0
        }
      }
    ])

{"$project": {"locations.datacenters.clusters.servers": 0}} will remove the servers field while keep all the other fields in the final output.

From the docs

If you specify the exclusion of a field other than _id, you cannot employ any other $project specification forms: i.e. if you exclude fields, you cannot also specify the inclusion of fields, reset the value of existing fields, or add new fields.

Reference: https://docs.mongodb.com/manual/reference/operator/aggregation/project/#exclude-fields

edited Oct 12, 2017 at 13:18

answered Oct 10, 2017 at 18:25

s7vr

76.2k12 gold badges118 silver badges135 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Paulo Pedroso Over a year ago

Thanks. It did gave me resources to figure out how to handle my requirement.

Alejandro · Accepted Answer · 2017-10-10 19:17:25Z

1

If you want to project only the clusters field for datacenters dc1.1 then below query will do it.

db.summary.find({"locations.datacenters": {$elemMatch: {"id": "dc1.1"}}},{"locations.datacenters.clusters":1, _id : 0})

Note, the _id:0 in projection is used to exclude the field _id to be displayed.

More information can be found at the following link: https://docs.mongodb.com/manual/reference/operator/projection/elemMatch/index.html

edited Oct 10, 2017 at 19:17

Alejandro

7,9274 gold badges40 silver badges70 bronze badges

answered Oct 10, 2017 at 17:09

Akhilesh Gupta

112 bronze badges

Comments

Rafiq · Accepted Answer · 2022-11-10 12:56:35Z

0

Nested Array, Push/Pull

 let doc = await Chat.findOneAndUpdate(
      {
        $and: [
          {
            _id: chatId,
          },
          {
            $or: [
              {
                user: req.user.id,
              },
              {
                'messages.$.recipients': {
                  $elemMatch: {
                    id: recipientId
                  },
                },
              },
            ],
          }
        ],
      },
      {
        $pull: {
          //@ts-ignore
          'messages.$.recipients': { id: recipientId },
        },
      },
      { new: true }
    );

answered Nov 10, 2022 at 12:56

Rafiq

12k5 gold badges44 silver badges46 bronze badges

Collectives™ on Stack Overflow

MongoDB $elemMatch projection on Nested Arrays

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related