Reputation: 7400
I have these documents:
{ "_id" : ObjectId("52abac78f8b13c1e6d05aeed"), "score" : 125494, "updated" : ISODate("2013-12-14T00:55:20.339Z"), "url" : "http://pictwittrer.com/1crfS1t" }
{ "_id" : ObjectId("52abac86f8b13c1e6d05af0f"), "score" : 123166, "updated" : ISODate("2013-12-14T00:55:34.354Z"), "url" : "http://bit.ly/JghJ1N" }
Now, i would like to get all documents sorted by this dynamic ranking:
ranking = score / (NOW - updated).abs
ranking is a float value where: - score is the value of scopre property of my document - the denominator is just the difference between NOW (when I'm executing this query) and updated field of my document
I'd want to do this because I want the old documents are sorted last
Upvotes: 0
Views: 2886
Reputation: 105
I'm new to Mongodb and aggregation frameworks but considering the answer Tim B gave I came up with this:
db.coll.aggregate(
{ $project : {
"ranking" : {
"$divide" : ["$score", {"$subtract":[new Date(), "$updated"]}]
}
}
},
{ $sort : {"ranking" : 1}})
Using $project you can reshape documents to insert precomputed values, in your case the ranking field. After that using $sort you can sort the documents by rank in the order you like by specifying 1 for ascending or -1 for descending.
I'm sorry for the terrible code formatting, I tried to make it as readable as possible.
Upvotes: 2
Reputation: 41188
Look at the MongoDB aggregation framework, you can do a project to create the score you want and then a sort to sort by that created score.
http://docs.mongodb.org/manual/core/aggregation-pipeline/
http://docs.mongodb.org/manual/reference/command/aggregate/#dbcmd.aggregate
Upvotes: 0