mongodb $lookup - but return from second collection only a part of the sub collection

Question

I have two collections:

• posts
• images

The images should be returned based on the post id(s) but only a part of the image sub collection should be returned. The result should only return the current entry of the sub collection that has the same lang.

That's what I have so far:

Post document

{
  "_id": 1,
  "lang": "en",
  "content": "foobar"
}

Image document

{
  "_id": ...,
  "post_id": 1,
  "title": [
    {
      "lang": "de",
      "value": "hallo welt"
    },
    {
      "lang": "en",
      "value": "hello world"
    }
  ],
}

Current code

Post.aggregate([
  { $match: query },
  {
    $lookup: {
      from: 'images',
      localField: '_id',
      foreignField: 'post_id',
      as: 'images',
    },
  },
  {
    $project: {
      content: 1,
      lang: 1,
      images: 1,
    },
  }
]);

But this will return the whole sub collection (en/de) but I just need the matched language. How can I solve this?

Thanks!

Answer 1

One way to limit what is returned from each document in"$lookup"'s "from" collection, is to use a "pipeline" in the "$lookup". Additionally, "$lookup"'s "let" optional field can be used to pass current document field values to the "pipeline".

db.posts.aggregate([
  {
    "$match": {
      // whatever filter is required goes here
      "_id": 1
    }
  },
  {
    "$lookup": {
      "from": "images",
      "localField": "_id",
      "foreignField": "post_id",
      "as": "images",
      // assign lang so it can be used in pipeline
      "let": {
        "lang": "$lang"
      },
      "pipeline": [
        {
          "$set": {
            "title": {
              // only return elements matching lang
              "$filter": {
                "input": "$title",
                "cond": {
                  "$eq": ["$$this.lang", "$$lang"]
                }
              }
            }
          }
        }
      ]
    }
  }
])

Try it on mongoplayground.net.