CODEWITHSUNDEEP
mongodb
Nathan Rutman
Nathan Rutman

Reputation: 2255

Get items of array by index in MongoDB

So I have a data structure in a Mongo collection (v. 4.0.18) that looks something like this…

{
  "_id": ObjectId("242kl4j2lk23423"),
  "name": "Doug",
  "kids": [
    {
      "name": "Alice",
      "age": 15,
    },
    {
      "name": "James",
      "age": 13,
    },
    {
      "name": "Michael",
      "age": 10,
    },
    {
      "name": "Sharon",
      "age": 8,
    }
  ]
}

In Mongo, how would I get back a projection of this object with only the first two kids? I want the output to look like this:

{
  "_id": ObjectId("242kl4j2lk23423"),
  "name": "Doug",
  "kids": [
    {
      "name": "Alice",
      "age": 15,
    },
    {
      "name": "James",
      "age": 13,
    }
  ]
}

It seems like I should easily be able to get them by index, but I'm not seeing anything in the docs about how to do that. The real-world problem I'm trying to solve has nothing to do with kids, and the array could be quite lengthy. I'm trying to break it up and process it in batches without having to load the whole thing into memory in my application.

EDIT (non-sequential indexes):

I noticed that since I asked about item 1 & 2 that $slice would suffice…however, what if I wanted items 1 & 3? Is there a way I can specify specific array indexes to return?

Any ideas or pointers for how to accomplish that?

Thanks!

Upvotes: 2

Views: 1474

Answers (1)

bauman.space
bauman.space

Reputation: 2023

You are looking for the $slice projection operator if the desired selection are near each other.

https://docs.mongodb.com/manual/reference/operator/projection/slice/

This would return the first 2

client.db.collection.find({"name":"Doug"}, { "kids": { "$slice": 2 } })

returns

{'_id': ObjectId('5f85f682a45e15af3a907f51'), 'name': 'Doug', 'kids': [{'name': 'Alice', 'age': 15}, {'name': 'James', 'age': 13}]}

this would skip the first kid and return the next two (second and third)

client.db.collection.find({"name":"Doug"}, { "kids": { "$slice": [1, 2] } })

returns

{'_id': ObjectId('5f85f682a45e15af3a907f51'), 'name': 'Doug', 'kids': [{'name': 'James', 'age': 13}, {'name': 'Michael', 'age': 10}]}

Edit:

Arbitrary selections 1 and 3 probably need to route through an aggregation pipeline rather than a simple query. The performance shouldn't be too much different assuming you have an index on the $match field.

Steps of your pipeline should be pretty obvious and you should be able to take it from here.

Hate to point to RTFM, but that's going to be super helpful here to at least be acquainted with the pipeline operations.

https://docs.mongodb.com/manual/reference/operator/aggregation/

Your pipeline should:

  1. $match on your desired query
  2. $set some new field kid_selection to element 1 (second element) and element 3 (4th element) since counting starts at 0. Notice the prefixed $ on the "kids" key name in the kid_selection setter. When referencing a key in the document you're working on, you need to prefix with $
  3. project the whole document, minus the original kids field that we've selected from
client.db.collection.aggregate([
    {"$match":{"name":"Doug"}},
    {"$set": {"kid_selection": [
        { "$arrayElemAt": [ "$kids", 1 ] }, 
        { "$arrayElemAt": [ "$kids", 3 ] }
    ]}},
    { "$project": { "kids": 0 } }
])

returns

{
  '_id': ObjectId('5f86038635649a988cdd2ade'), 
  'name': 'Doug', 
  'kid_selection': [
     {'name': 'James', 'age': 13}, 
     {'name': 'Sharon', 'age': 8}
   ]
}

Upvotes: 3

Related Questions