Sum of All Messages Sent - mongodb

Question

Mongodb Version 2.6.9
I'm attempting to count the total a company has been involved in a messaging interaction. I'm able to get one side of the interaction using the aggregate $group, but I've come up empty on essentially looking at the two fields and aggregating those together for each unique company ID.
The sender_id and receiver_id relate to the same company Id's.
{ "_id" : a, "sender_id" : 1, "receiver_id" : 2, payload: "data" }
{ "_id" : b, "sender_id" : 2, "receiver_id" : 5, payload: "data" }
{ "_id" : c, "sender_id" : 2, "receiver_id" : 4, payload: "data" }
{ "_id" : d, "sender_id" : 3, "receiver_id" : 2, payload: "data" }
{ "_id" : e, "sender_id" : 4, "receiver_id" : 1, payload: "data" }
Using the above data structure, I attempting to produce a result set similar to
{ "_id" : 1, count: 2}
{ "_id" : 2, count: 4}
{ "_id" : 3, count: 1}
{ "_id" : 4, count: 2}
{ "_id" : 5, count: 1}
where for example Company 2 was involved in messages a, b, c, d.

Answer 1

Your options are limited in 2.6 pipeline. You can try below pipeline.
$group with $push to create single value array for both sender_id and receiver_id.
$project with $setUnion to merge ids into single array.
$unwind and $group to count the occurrences.
db.collection.aggregate({
"$group": {
"_id": "$_id",
"sender_id": {
"$push": "$sender_id"
},
"receiver_id": {
"$push": "$receiver_id"
}
}
}, {
"$project": {
"id": {
"$setUnion": ["$sender_id", "$receiver_id"]
}
}
}, {
"$unwind": "$id"
}, {
"$group": {
"_id": "$id",
"count": {
"$sum": 1
}
}
})
You can use below pipeline for newer versions. Use [] to create array.
db.collection.aggregate({
$project: {
id: ["$sender_id", "$receiver_id"]
}
}, {
$unwind: "$id"
}, {
$group: {
_id: "$id",
count: {
$sum: 1
}
}
})