Getting data from subdocument between dates - mongodb

Question

I'm trying to get a certain amount of historical data from my documents similar to the SQL's Between clause. I searched for similar questions and managed to find this query but it only returns the first element of the array, so it's not enough:
db.docs.find({ _id: "eraj4983tj3" }, {history: { $elemMatch : {time: {$gt: new ISODate("2016-03-21T20:53:33.662Z") , $lt: new ISODate("2016-03-21T20:54:20.313Z") } } } });
My documents look like this:
{
"_id": "eraj4983tj3",
"descr": "somestuff",
"history": [
{
"time": ISODate("2016-03-21T20:52:31.389Z"),
"value": 103.91
},
{
"time": ISODate("2016-03-21T20:53:33.663Z"),
"value": 106.91
},
{
"time": ISODate("2016-03-21T20:53:45.179Z"),
"value": 104.91
},
{
"time": ISODate("2016-03-21T20:54:20.313Z"),
"value": 105.11
},
{
"time": ISODate("2016-03-21T20:54:53.649Z"),
"value": 105.41
},
{
"time": ISODate("2016-03-21T20:55:12.998Z"),
"value": 115.91
}
]
}
And the result of my query should return this:
{
"_id": "eraj4983tj3",
"history": [
{
"time": ISODate("2016-03-21T20:53:45.179Z"),
"value": 104.91
},
{
"time": ISODate("2016-03-21T20:54:20.313Z"),
"value": 105.11
},
{
"time": ISODate("2016-03-21T20:54:53.649Z"),
"value": 105.41
}
]
}
How should i write my query to achieve this result?

Answer 1

What you want is not possible with a simple query. However, you can use an aggregation:
db.yourColl.aggregate([
{ $match:{ _id:"eraj4983tj3" } },
{ $unwind: "$history" },
{ $match:{ "history.time":{
$gt: new ISODate("2016-03-21T20:53:33.662Z") ,
$lt: new ISODate("2016-03-21T20:54:20.313Z")
}}},
{ $group:{ _id:"$_id", history:{ $push:"$history" }}}
])
Which gives you the (correct) result of
{
"_id" : "eraj4983tj3",
"history" : [
{ "time" : ISODate("2016-03-21T20:53:33.663Z"), "value" : 106.91 },
{ "time" : ISODate("2016-03-21T20:53:45.179Z"), "value" : 104.91 }
]
}
Now, let us examine the stages:
{ $match:{ _id:"eraj4983tj3" } }: Find the correct doc
{ $unwind: "$history" } Create a new document for each array item and send it down the pipeline.
The output of this stage looks like this:
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:52:31.389Z"), "value" : 103.91 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:53:33.663Z"), "value" : 106.91 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:53:45.179Z"), "value" : 104.91 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:54:20.313Z"), "value" : 105.11 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:54:53.649Z"), "value" : 105.41 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:55:12.998Z"), "value" : 115.91 } }
{ $match:{ "history.time":{ $gt: new ISODate("2016-03-21T20:53:33.662Z"), $lt: new ISODate("2016-03-21T20:54:20.313Z")}}} We match again to find the history items from above stage that match our date range. The output of this stage looks like this:
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:53:33.663Z"), "value" : 106.91 } }
{ "_id" : "eraj4983tj3", "descr" : "somestuff", "history" : { "time" : ISODate("2016-03-21T20:53:45.179Z"), "value" : 104.91 } }
{ $group:{ _id:"$_id", history:{ $push:"$history" }}} We group by _id (nothing new here), and for each document of the output of above stage, we $push its value for history (an object) to the array history in the newly created grouped document.

Answer 2

The positional operator $ and the elemMatch will only return one sub document.