MongoDB+Scala: Accessing deep nested data

Question

I think there should be an easy solution around, but I wasn't able to find it.

I start accessing data from MongoDB with the following in Scala:

val search = MongoDBObject("_id" -> new ObjectId("xxx"))
val fields = MongoDBObject("community.member.name" -> 1, "community.member.age" -> 1)

for (res <- mongoColl.find(search, fields)) {
    var memberInfo = res.getAs[BasicDBObject]("community").get
    println(memberInfo)
}

and get a BasicDBObject as result:

{
"member" : [
    {
        "name" : "John Doe",
        "age" : "32",
    },{
        "name" : "Jane Doe",
        "age" : "29",
    },
    ...
]
}

I know that I can access values with getAs[String], though this is not working here... Anyone has an idea? Searching for a solution for several hours...

Sergey Passichenko · Answer 1 · 2013-03-15T09:46:57.480

2

If you working with complex MongoDB objects, you can use Salat, which provides simple case class serialization.
Sample with your data:

case class Community(members:Seq[Member], _id: ObjectId = new ObjectId)

case class Member(name:String, age:Int)

val mongoColl: MongoCollection = _

val dao = new SalatDAO[Community, ObjectId](mongoColl) {}

val community = Community(Seq(Member("John Doe", 32), Member("Jane Doe", 29)))
dao.save(community)

for {
    c <- dao.findOneById(community._id)
    m <- c.members
} println("%s (%s)" format (m.name, m.age))

edited Mar 15 '13 at 09:46

answered Mar 15 '13 at 07:00

Sergey Passichenko

6,920
1
28
29

Sounds interesting. Could you provide a short example adressing deeper nested data? – Max Leeds Mar 15 '13 at 09:12
This looks very promising and clearly structured. In the long term I will go for it. But to solve my urgent problem the answer from anoopelias is short and fine. I would like to upvote your post but I have to get some reputation first... will make a task for that later on! :-) Thanks again! – Max Leeds Mar 15 '13 at 13:23

score 0 · Accepted Answer · answered Mar 15 '13 at 06:50

0

I think you should try

val member = memberInfo.as[MongoDBList]("member").as[BasicDBObject](0)
println(member("name"))

answered Mar 15 '13 at 06:50

anoopelias

9,240
7
26
39

I am suspecting there is a less verbose solution. You can try `memberInfo.as[MongoDBList]("member").as(0)` and see if it can infer the type. – anoopelias Mar 16 '13 at 09:33
Also yes, salat is they way to go in the long run. It is an elegant wrapper over casbah. – anoopelias Mar 16 '13 at 09:36

score 0 · Answer 3 · answered Mar 15 '13 at 08:17

This problem has not to do really with MongoDB, but rather with your data structure. Your JSON/BSON data structure includes

An object community, which includes
- An array of members
  - Each member has properties name or age.

Your problem is completely equivalent to the following:

case class Community(members:List[Member])

case class Member(name:String, age:Int)

val a = List(member1,member2)

// a.name does not compile, name is a property defined on a member, not on the list

Luciano · Answer 4 · 2015-03-18T17:18:02.007

Yes, you can do this beautifully with comprehensions. You could do the following:

for { record <- mongoColl.find(search,fields).toList
      community <- record.getAs[MongoDBObject]("community")
      member <- record.getAs[MongoDBObject]("member")
      name <- member.getAs[String]("name") } yield name

This would work just to get the name. To get multiple values, I think you would do:

for { record <- mongoColl.find(search,fields).toList
      community <- record.getAs[MongoDBObject]("community")
      member <- record.getAs[MongoDBObject]("member")
      field <- List("name","age") } yield member.get(field).toString

MongoDB+Scala: Accessing deep nested data

4 Answers4