Need to aggregate by hour and $avg not recognized - mongodb

Question

From a MongoDB collection storing data with time stamps I need to return a single record for each hour.
So far I have selected the set of records between two dates successfully, but I cant figure how to build the hourly record I need in the $group clause.
var myName = "CollectionName"
//schema for mongoose
var mySchema = new Schema({
dt: Date,
value: Number
});
var myDB = mongoose.createConnection('mongodb://localhost:27017/MYDB');
myDBObj = myDB.model(myName, evalSchema, myName);
The match in this aggregate call works fine, and the $hour creates a record for each hour in the day.. but I don't know how to recreate the a full date and get an error "unknown group operator $avg" ...
myDBObj.aggregate([
{
$match: { "dt": { $gt: new Date("October 13, 2010 12:00:00"), $lt: new Date("November 13, 2010 12:00:00") } }
},{
$group: {
"_id": { "dt": { "$hour": "$dt" } , "price": { "$avg": "$price" }}
}], function (err, data) { if (err) { return next(err); } res.json(data); });
I think I need to use $dayOfYear so there is different records for each hour of each day, and include a new Date() somewhere ...
Can someone help me do this correctly? any help is appreciated.

Answer 1

The $group pipeline stage works by "grouping" all data by the "key" specified for _id. Other fields you are actually aggregating are separate from the _id value and are their own field properties.
So your $group becomes this instead:
{ "$group": {
"_id": { "$hour": "$dt" },
"price": { "$avg": "$price" }
}}
Or if you want that broken by day then make a compound key:
{ "$group": {
"_id": {
"day": { "$dayOfYear": "$dt" },
"hour": { "$hour": "$dt" }
},
"price": { "$avg": "$price" }
}}
Or just use date math to produce Date objects rounded by hour:
{ "$group": {
"_id": {
"$add": [
{ "$subtract": [
{ "$subtract": [ "$dt", new Date(0) ] },
{ "$mod": [
{ "$subtract": [ "$dt", new Date(0) ] },
1000 * 60 *60
]}
]},
new Date(0)
]
},
"price": { "$avg": "$price" }
}}
Where subrtacting another date object (epoch date) from another prodces a numeric value you can round ( 1000 milliseconds, 60 seconds, 60 minutes = 1 hour ) with the applied math, and adding a number to a date object produces a date corresponding to that value.
So your problem was you had everything in the _id, where the $avg accumulator is not recognised. All accumulators need to be specified outside of the grouping key. That is the intent.
If you want to make an accumulator value part of a grouping key ( does not seem relevant here though ), you instead follow with another group stage, referencing the field that was produced from the former.