CODEWITHSUNDEEP
mongodbmongodb-queryaggregation-framework
Rafael Herculano
Rafael Herculano

Reputation: 75

Count Array Elements Matching Condition

I have a collection in MongoDB and their objects look like this:

Object 1

{
    "_id" : ObjectId("5afde62a91952a2980a1b751"),   
    "notificationName" : "Teste Agendamento Pontual",
    "messages" : [ 
        {
            "timestamp" : ISODate("2018-05-17T20:29:33.045Z"),            
            "message" : "Teste Agendamento Pontual"
        }, 
        {
            "timestamp" : ISODate("2018-05-17T20:29:33.051Z"),            
            "message" : "Teste"
        }, 
        {
            "timestamp" : ISODate("2018-05-17T20:29:44.680Z"),            
            "message" : "OK"
        }        
    ]
}

Object 2

{
    "_id" : ObjectId("5afde62a9194322980a1b751"),   
    "notificationName" : "Teste Agendamento Pontual",
    "messages" : [ 
        {
            "timestamp" : ISODate("2018-05-17T20:29:33.045Z"),            
            "message" : "Teste Agendamento Pontual"
        }, 
        {
            "timestamp" : ISODate("2018-05-17T20:29:33.051Z"),            
            "message" : "NOT OK"
        }, 
        {
            "timestamp" : ISODate("2018-05-17T20:29:44.680Z"),            
            "message" : "asdsadasd"
        }        
    ]
}

...

I'm trying to get a result grouped by notificationName with a count of objects with OK messages and other count with objects with NOT OK . I think I need to make a $cond in $group to check message property, but I'm not sure about that. And not sure if thats the best way to achieve this.

Upvotes: 0

Views: 168

Answers (1)

Neil Lunn
Neil Lunn

Reputation: 151170

If you want matching array elements then you want $filter instead and count them using $size:

db.collection.aggregate([
  { "$group": {
    "_id": "$notificationName",
    "countOk": {
      "$sum": {
        "$size": {
          "$filter": {
            "input": "$messages.message",
            "cond": { "$eq": [ "$$this", "OK" ] }
          }
        } 
      }
    }
  }}
])

The $filter has it's own cond argument which is a logical condition to return a boolean value determining whether the array element matches that condition or not and can be returned. Since this would only then return an array of values from message where the value is "OK" using the $eq comparison operator to test, then you "count" the resulting array members using $size.

Because you are "grouping" you use $group as the stage to do this, and because you are "accumulating" you use the $sum operator to "add up" all the returned $size results from each document sharing the same grouping key

Upvotes: 4

Related Questions