I have the following:
{_id: blah, anotherId: "externalId", description: "foo"},
{_id: blah, anotherId: "externalId", description: "bar"},
...
I want:
{_id: blah, anotherId: "externalId", descriptions: ["foo", "bar"]}
I know I can simply write this but it's very slow as I have millions of records.
db.collectionOfAnotherId.find().forEach(function(r){
var x = {anotherId: r.id, descriptions: []};
db.myCollection.find({anotherId: x.anotherId}).forEach(function(d){
x.descriptions.push(d.description); });
db.newCollection.save(x);
})
Any ideas?
You could use the aggregation framework for this. For example, consider using the $lookup operator to do a left join to myCollection and writing the results of the following $group pipeline to the new collection using $push to create the descriptions array as:
db.collectionOfAnotherId.aggregate([
{
"$lookup": {
"from": "myCollection",
"localField": "id",
"foreignField": "anotherId",
"as": "d"
}
},
{ "$unwind": "$d" },
{
"$group": {
"_id": "$_id",
"anotherId": { "$first": "$id" },
"descriptions": { "$push": "$d.description" }
}
},
{ "$out": "newCollection" }
])