Searching parent id in cloudant - couchdb

Question

I have a Cloudant DB with the following structure:
{id: 1, resource:”john doe”, manager: “john smith”, amount: 13}
{id: 2, resource:”mary doe”, manager: “john smith”, amount: 3}
{id: 3, resource:”john smith”, manager: “peter doe”, amount: 10}
I needed a query to return the sum of amount, so I've built a query with emit(doc.manager, doc.amount) which returns
{"rows":[
{"key":"john smith","value":16},
{"key":"peter doe","value":10}]}
It is working like a charm. However I need the manager ID along with Manager name. The result I am looking for is:
{"rows":[
{"key":{"john smith",3},"value":16},
{"key":{"peter doe",null},"value":10}]}
How should I build a map view to search the parent ID?
Thanks,
Erik

Answer 1

Unfortunately I don't think there's a way to do exactly what you want in one query. Assuming you have the following three documents in your database:
{
"_id": "1",
"resource": "john doe",
"manager": "john smith",
"amount": 13
}
--
{
"_id": "2",
"resource": "mary doe",
"manager": "john smith",
"amount": 3
}
--
{
"_id": "3",
"resource": "john smith",
"manager": "peter doe",
"amount": 10
}
The closest thing to what you want would be the following map function (which uses a compound key) and a _sum reduce:
function(doc) {
emit([doc.manager, doc._id], doc.amount);
}
This would give you the following results with reduce=false:
{"total_rows":3,"offset":0,"rows":[
{"id":"1","key":["john smith","1"],"value":13},
{"id":"2","key":["john smith","2"],"value":3},
{"id":"3","key":["peter doe","3"],"value":10}
]}
With reduce=true and group_level=1, you essentially get the same results as what you already have:
{"rows":[
{"key":["john smith"],"value":16},
{"key":["peter doe"],"value":10}
]}
If you instead do reduce=true and group=true (exact grouping) then you get the following results:
{"rows":[
{"key":["john smith","1"],"value":13},
{"key":["john smith","2"],"value":3},
{"key":["peter doe","3"],"value":10}
]}
Each unique combination of the manager and _id field is summed, which unfortunately doesn't give you what you want. To accomplish what you want to accomplish, I think your best but would be to sum up the values after querying the database.