Sum results of multiple grouped documents - couchbase

Question

I have setup a schema where I associate two different entities: Users and Brands. This association is stored in the document UserBrand. A user may have multiple brands, and in each UserBrand document, the information there contained is different.
Here is an example of three documents:
{
type: "UserBrand",
userId: "x",
brandId: 1,
value: 100
}
{
type: "UserBrand",
userId: "x",
brandId: 2,
value: 50
}
{
type: "UserBrand",
userId: "y",
brandId: 1,
value: 150
}
As you see, user 'x' follows multiple brands. How can I get the sum of the values grouped by user, while fetching the user information as well?
So far I am able to join both document types (User and UserBrand) through this query:
SELECT ub.*, u.name FROM bucket as ub JOIN bucket as u ON KEYS "User_" || ub.userId WHERE ub.type="UserBrand" AND u.type="User" AND (ub.brand=1 OR ub.brand=2)
but I cannot seem to group the records and sum the values for a single user. Is it possible to do so on a single query?
If possible, the end result should be something like:
{
type: "UserBrand",
name: "name1',
userId: "x",
brandId: 1,
value: 150
}
{
type: "UserBrand",
name: "name2',
userId: "y",
brandId: 1,
value: 150
}

Answer 1

Why are you using 'join' instead of simply 'group by' and aggregation? If I wanted to sum the values for each distinct userId, I would write a query as follows:
select userId, sum(value) from bucket group by userId;
Note that when using 'group by', every field in the select list must be either part of the 'group by' list (userId above) or an aggregation expression (like sum(value) above). But what if you want to include the field 'name' in the result? If you know that 'name' always has the same value for a given id, you could use an aggregation function like 'max' or 'min' to bring up the value (since max of a set of identical values will be any of the values):
select userId, max(name), sum(value) from bucket group by userId;
For a field like 'brandId', which seems to have different values, you'd want to use an aggregation function like 'array_agg', which will take all the different values of 'brandId' and put them in an array. E.g.,
select userId, max(name), sum(value), array_agg(distinct brandId) from bucket group by userId;