-
Type:
Improvement
-
Resolution: Done
-
Priority:
Major - P3
-
None
-
Affects Version/s: 2.1.2
-
Component/s: Aggregation Framework
-
Query
-
None
-
3
-
None
-
None
-
None
-
None
-
None
-
None
here's a list of documents:
{_id:1, manufacturer: "Sony", features: ["fast", "sleek", "sexy"]}
{_id:2, manufacturer: "Sony", features: ["lightweight", "sleek"]}
{_id:3, manufacturer: "Sony", features: ["cheap", "fast"]}
{_id:4, manufacturer: "Apple", features: ["fast", "sleek", "usable"]}
{_id:5, manufacturer: "Apple", features: ["lightweight", "sleek", "cheap"]}
{_id:6, manufacturer: "Apple", features: ["fast", "usable", "codecs"]}
I'd like to output to look like:
{
"manufacturer": {
"Sony":3,
"Apple":3
},
"features": {
"fast": 4,
"cheap": 2,
"codecs": 1,
"usable": 2,
"lightweight": 2,
"sleek": 3,
"sexy": 1,
}
}
Here's my opinion on how to make this request:
db.coll.aggregate(
{$count: ["manufacturer", "features"]}
)
Alternately, I'd be fine with a new group aggregation function, like so:
db.coll.aggregate({
$group: {
manufacturer: {$count: "$manufacturer"},
features: {$count: "$features"}
}
})
Although this obviously requires that the need to specify an _id be lifted, or accounted for in some way.
- is related to
-
SERVER-12927 Support ROLLUP function in the $group aggregation operator
-
- Backlog
-
-
-
- Closed
-