MongoDB Aggregation Function Returning Undefined - mongodb

Question

I'm attempting to use use the new MongoDB aggregation features to tally some statistics by date. Below is a sample of the documents that I am working with, my attempted code and desired result. The aggregation function retuns "UNDEFINED". Can someone tell me why that is? And secondly, I want my aggregation function to group results by date in mm-dd-yyyy format. However as it is currently written I think the code is going to execute the aggregation by the full ISO date. Can someone please tell me how to fix this?
DOCUMENT EXAMPLE
{
user: "2A8761E4-C13A-470E-A759-91432D61B6AF-25982-0000352D853511AF",
language: "English",
imageFileName: "F7A5ED9-D43C-4671-A5C6-F06C7E41F902-7758-000008371FB5B834",
audioFileName: "F6D5727D-9377-4092-A28A-AA900F02653D-7758-0000083749066CF2",
date: ISODate("2012-10-22T02:43:52Z"),
correct: "1",
_id: ObjectId("5084b2e8179c41cc15000001")
}
AGGREGATION FUNCTION
var getUserStats = function(user, language, callback) {
var guessCollection = db.collection('Guesses');
guessCollection.aggregate(
{ $match: {
user: user,
language: language,
}},
{ $sort: {
date: 1
}},
{ $project : {
user : 1,
language : 1,
date : 1,
correct : 1,
incorrect : 1,
} },
{ $unwind : "$language" },
{ $group : {
_id : "$date",
correct : { $sum : "$correct" },
incorrect : { $sum : "$incorrect" }
} }
, function(err, result){
console.log(result);
callback(result);
});
DESIRED RESULT
{
"result" : [
//...snip...
{
"_id" : "2A8761E4-C13A-470E-A759-91432D61B6AF-25982-0000352D853511AF",
"correct" : 32,
"incorrect" : 17,
"date" : 2012-10-22
},
{
"_id" : "2A8761E4-C13A-470E-A759-91432D61B6AF-25982-0000352D853511AF",
"correct" : 16,
"incorrect" : 7,
"date" : 2012-10-23
}
],
"Ok" : 1
}

Answer 1

Regarding your first question about it returning undefined, there are two problems:
You are using the $unwind operator on a field ($language) that isn't an array.
You are using the $sum operator on a string field ($correct); that's only supported for number fields.
For your second question about grouping on just the date, you need to project the date components you want to group on and then use those components in your $group operator's _id value:
For example:
test.aggregate(
{ $match: {
user: user,
language: language
}},
{ $sort: {
date: 1
}},
{ $project : {
user : 1,
language : 1,
year : { $year: '$date' },
month : { $month: '$date' },
day : { $dayOfMonth: '$date'},
correct : 1,
incorrect : 1
}},
{ $group : {
_id : { year: "$year", month: "$month", day: "$day" },
correct : { $sum : "$correct" },
incorrect : { $sum : "$incorrect" }
}},
function(err, result){
console.log(result);
}
);
Produces output of:
[ { _id: { year: 2012, month: 10, day: 22 },
correct: 0,
incorrect: 0 } ]
You can assemble that into '2012-10-22' in code from there.