Question

I have two collections. “people” are connected with a “location” like so:

location_id = ObjectId()

db.people.insert(
{
  _id : ObjectId(),
  name : "Nick",
  location : location_id
});

db.locations.insert(
{
 _id : location_id,
  city : "Cape Town"
});

I would like to create a histogram of locations giving the count of people in each city. But I can’t seem to do it with the Mongo group command because they are different collections. Is the correct way to do it with map/reduce?

Answer 1

map-reduce also won’t help you here. It also operates on a single collection.

It seems to me that you’re trying to apply your relational design skills here. Don’t do that. MongoDB requires a fresh approach.

If I were you, I’d probably denormalize location into people collection.

db.people.insert(
{
  _id : ObjectId(),
  name : "Nick",
  location : {city: "Cape Town"}
});

This way it’s possible to analyze this data with map-reduce, because everything is in the same collection.