MongoDB - count instances across multiple documents

Question

I've got around 100 documents that look like these:

{
    "_id": "p1",
    "when": "2016-06-01T00:00:00.000Z",
    "users": [
      "u1",
      "u2"
    ]
},

{
    "_id": "p2",
    "when": "2016-06-01T00:00:00.000Z",
    "users": [
      "u1",
      "u3"
    ]
},

{
    "_id": "p3",
    "when": "2016-06-01T00:00:00.000Z",
    "users": [
      "u1",
      "u2",
      "u3"
    ]
}

I want to run a query that will return all the user ids that exist in exactly 3 documents between two dates.

So for example, using the above data and for the dates 2016-04-01 to 2016-07-01 it would return u1 (u1 exists in 3 records between those dates).

I presume it'll use some sort of group by or count and the aggregation framework, but I'm looking at the most performant / efficient way of doing it so I'm open to suggestions. Bear in mind that although there may only be 100 documents there could be 1 million users (so the user arrays in each document could be huge).

Any thoughts?

Answer 1

you can try this query

db.getCollection('collectionName').aggregate([
    {$match:{$and:[{"when": { "$gte": '2016-06-01T00:00:00.000Z' }},{"when": { "$lte": '2016-06-01T00:00:00.000Z' }}]}},
    {$unwind: "$users"},
    {$group:{_id:"$users", count:{$sum:1}}},
    {$match:{"count":{"$eq":3}}}
])

in this query I have followed some steps

1. Filter record between dates
2. Count how many times exist each user
3. Filter to get exact 3 times exist users