MongoDB Group by field and show array of grouped items? - mongodb

Question

I have a collection of Projects in where projects are like this:
{
"_id" : ObjectId("57e3e55c638cb8b971"),
"allocInvestor" : "Example Investor",
"fieldFoo" : "foo bar",
"name" : "GTP 3 (Roof Lease)"
}
I want to receive a list of projects grouped by allocInvestor field and only show fields: name and id
If I use aggregate and $group like this:
db.getCollection('projects').aggregate([
{"$group" : {
_id:"$allocInvestor", count:{$sum:1}
}
}
])
I receive a count of project per allocInvestor but I need is to receive the list of allocInvestor with subarray of projects per allocInvestor.
I'm using meteor by the way in case that helps. But I first want to get the query right on mongodb then try for meteor.

Answer 1

You can use $push or $addToSet to create a list of name and _id per every group.
$push allows duplicates and $addToSet does not add an element to the list again, if it is already there.
db.getCollection('projects').aggregate([
{ "$group" : { _id : "$allocInvestor",
count : {$sum : 1},
"idList" : {"$addToSet" : "$_id"},
"nameList" : {"$addToSet":"$name"}
}
}
]);
To get the name and _id data in a single list:
db.getCollection('projects').aggregate([
{ "$group" : { _id : "$allocInvestor", "projects" : {"$addToSet" : {id : "$_id", name: "$name"}}}},
{"$project" : {"_id" : 0, allocInvestor : "$_id", "projects" : 1 }}
]);

Answer 2

Use the $$ROOT operator to reference the entire document and then use project to eliminate the fields that you do not require.
db.projects.aggregate([
{"$group" : {
"_id":"$allocInvestor",
"projects" : {"$addToSet" : "$$ROOT"}
}
},
{"$project" : {
"_id":0,
"allocInvestor":"$_id",
"projects._id":1
"projects.name":1
}
}
])