Mongodb get last combination in aggregation framew

has a collection of messages:

{
  "date": NumberLong(1421134514),
  "sender": "53172480f9cd0e682840b9f7",
  "recipient": "52f37fbaf9cd0e02773c97b1",
  "isRead": false,
  "_id": "54b4cab2f6a48ce34f8b5a75",
  "text": "Hello!"
},
{
  "date": NumberLong(1421135561),
  "sender": "53172480f9cd0e682840b9f7",
  "recipient": "52f37fbaf9cd0e02773c97b1",
  "isRead": false,
  "_id": "54b4cec9f6a48ce34f8b6429",
  "text": "Hello 2!"
},
{
  "date": NumberLong(1421135618),
  "sender": "53072122f9cd0ee76306dc5a",
  "recipient": "52f37fbaf9cd0e02773c97b1",
  "isRead": false,
  "_id": "54b4cf02f6a48ce54f8b62f9",
  "text": "Hello 3!"
},
{
  "date": NumberLong(1421136457),
  "sender": "52f37fbaf9cd0e02773c97b1",
  "recipient": "52ea178ff9cd0e9f24d776b4",
  "isRead": false,
  "_id": "54b4d249f6a48ce54f8b6b9f"
  "text": "Hello 4!"
}

It is necessary to choose the latter dialogues, where the current user is either the sender or recipient.

For example, for a user with ID = '52f37fbaf9cd0e02773c97b1', should get 3 records.

It turns out to build two separate queries:

$result = \DB::$connection->message->aggregate(array(
    array('$match' => array('sender' => \Core::getModule('users')->user->_id)),
    array('$group' => array('_id' => '$recipient')),
));

$result2 = \DB::$connection->message->aggregate(array(
    array('$match' => array('recipient' => \Core::getModule('users')->user->_id)), 
    array('$group' => array('_id' => '$sender')),
));

Is it possible to somehow combine these two queries into one and sort the records by date?

What you need here is a "unique key" value generated for each combination of "sender and recipient". If you want to do this on a regular basis then I would suggest storing the value in the document. But this is a way you can get out of the problem using the aggregation framework:

 db.messages.aggregate([
    { "$project": {
       "combined": { "$map": {
           "input": { "$literal": ["A","B"] },
           "as": "bin",
           "in": { "$cond": [ 
               { "$eq": [ "$$bin", "A" ] },
               "$sender",
               "$recipient"
           ]}
       }},
       "doc": "$$ROOT"
    }},
    { "$unwind": "$combined" },
    { "$sort": { "_id": 1, "combined": 1, "doc.date": -1 } },
    { "$group": {
        "_id": "$_id",
        "combined": { "$push": "$combined" },
        "doc": { "$first": "$doc" }
    }},
    { "$group": {
        "_id": "$combined",
        "doc": { "$first": "$doc" }
    }}
])

That narrows down your sample to the "unique" 3 combinations on "sender / recipient" and returns the last document in the conversation between that "pair":

{
    "_id" : [
            "52f37fbaf9cd0e02773c97b1",
            "53172480f9cd0e682840b9f7"
    ],
    "doc" : {
            "_id" : "54b4cec9f6a48ce34f8b6429",
            "date" : NumberLong(1421135561),
            "sender" : "53172480f9cd0e682840b9f7",
            "recipient" : "52f37fbaf9cd0e02773c97b1",
            "isRead" : false,
            "text" : "Hello 2!"
    }
}
{
    "_id" : [
            "52f37fbaf9cd0e02773c97b1",
            "53072122f9cd0ee76306dc5a"
    ],
    "doc" : {
            "_id" : "54b4cf02f6a48ce54f8b62f9",
            "date" : NumberLong(1421135618),
            "sender" : "53072122f9cd0ee76306dc5a",
            "recipient" : "52f37fbaf9cd0e02773c97b1",
            "isRead" : false,
            "text" : "Hello 3!"
    }
}
{
    "_id" : [
            "52ea178ff9cd0e9f24d776b4",
            "52f37fbaf9cd0e02773c97b1"
    ],
    "doc" : {
            "_id" : "54b4d249f6a48ce54f8b6b9f",
            "date" : NumberLong(1421136457),
            "sender" : "52f37fbaf9cd0e02773c97b1",
            "recipient" : "52ea178ff9cd0e9f24d776b4",
            "isRead" : false,
            "text" : "Hello 4!"
    }
}