The number of readers per controller can have an impact on overall system performance. Determining the number of readers per controller is a great application for the aggregation pipeline, and keepcli

We’ll create a new file to hold our aggregation pipeline query called readersPerController.json.

[
    {}
]

The first operation is to include in the pipeline only MercuryReader objects. The type of object is identified in the _t property so the $match operator will be:

{
    "$match":{"_t":"MercuryReader"}
}

The controller that a reader is attached to is identified by a reference in the ObjectLinks array marked with a Relation value of Controller. Therefore we need to use $unwind operator on the ObjectLinks array and then filter only the Controller ObjectLinks

{
    "$unwind":"$ObjectLinks"
},
{
    "$match":{"ObjectLinks.Relation":"Controller"}
}

The final two stages of the pipeline will be $group operators. The first group will count the number of readers per controller:

{
    "$group":{"_id":"$ObjectLinks.LinkedObjectId","count":{"$sum":1}}
}

Now our pipeline looks something like this, showing the key value of the controller, and the number of readers attached.

{"_id":{"$oid":"5d30868833442c01a4a9eedd"},"count":1}
{"_id":{"$oid":"57e036ac92dacd0d6cc6b480"},"count":6}

Finally, value that we’re interested in for this exercise is the average number of readers per controller, and so a final $group operator is required, this time using null as the grouping value, thereby grouping the entire pipeline in a single group.

{
    "$group":{"_id":null, "average":{"$avg":"$count"}}
}

The completed aggregation pipeline file should look something like this:

[
    {
        "$match": {
            "_t": "MercuryReader"
        }
    },
    {
        "$unwind": "$ObjectLinks"
    },
    {
        "$match": {
            "ObjectLinks.Relation": "Controller"
        }
    },
    {
        "$group": {
            "_id": "$ObjectLinks.LinkedObjectId",
            "count": {
                "$sum": 1
            }
        }
    },
    {
        "$group": {
            "_id": null,
            "average": {
                "$avg": "$count"
            }
        }
    }
]

Use keepcli to execute this aggregation pipeline passing the output to jq for easy formatting

keepcli aggregate < readersPerController.json | jq .