Search code examples
jsonscalajsonpathplay-json

Scala - How to Split all List of List Json Nodes using json-path

I have a Json from which I want to pick List of List Json, where instance can be multiple inside List. Using json-path easily we can pick if giving index number of List/Array. But in a Big File we don't know total how many instance will be there and we have not to loose any data. So number of instance has to be check in a dynamic way and pick seperate json for all inside List node. Additionally has to create relation_path also for all the Data.

Can Anyone suggest How to check if a json node is Array/List (Ex : 2 Drive) and how many nested List objects are available like 2 Partition in 1st Drive and 1 Partition in 2nd Drive. These numbers are not fixed to be provide in json-path code.

Input List of List Json :

{"Start":{"HInfo":{"InfoId":"650FEC74","Revision":"5.2.0.51","Drive":[{"InfoId":"650FEC74","Index":0,"Name":"Drive0","Partition":[{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":0},{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":1}]},{"InfoId":"650FEC74","Index":1,"Name":"Drive1","Partition":{"InfoId":"650FEC74","DriveID":"3F275869","Index":0}}]}}}

Output List of Json :

[{"Partition":[{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":0},{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":1}],"relation_tree":"Start/HInfo/Drive/Drive-1/Partition"},{"Partition":{"InfoId":"650FEC74","DriveID":"3F275869","Index":0},"relation_tree":"Start/HInfo/Drive/Drive-2/Partition"}]

What I am trying using json-path, but this is not fittable as I here I am providing Index Number manually, which is not possible in all the case as index number can be 0 to any.

val jsonString = """{"Start":{"HInfo":{"InfoId":"650FEC74","Revision":"5.2.0.51","Drive":[{"InfoId":"650FEC74","Index":0,"Name":"Drive0","Partition":[{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":0},{"InfoId":"650FEC74","DriveID":"F91B1F36","Index":1}]},{"InfoId":"650FEC74","Index":1,"Name":"Drive1","Partition":{"InfoId":"650FEC74","DriveID":"3F275869","Index":0}}]}}}"""
  val jsonStr: JsValue = Json.parse(jsonString)
    var pruneJson1 = (__ \ "Partition").json.copyFrom((__ \ "Start" \ "HInfo" \ "Drive" \ (0) \ "Partition").json.pick)
    val finalPartitionPrune1 = Option(jsonStr.transform(pruneJson1)).get.get.as[JsObject] + ("relation_tree" -> Json.toJson("Start"+"/"+"HInfo"+"/"+"Drive"+"/"+"Drive-1"+"/"+"Partition"))

    println(finalPartitionPrune1)

  var pruneJson2 = (__ \ "Partition").json.copyFrom((__ \ "Start" \ "HInfo" \ "Drive" \ (1) \ "Partition").json.pick)
  val finalPartitionPrune2 = Option(jsonStr.transform(pruneJson2)).get.get.as[JsObject] + ("relation_tree" -> Json.toJson("Start"+"/"+"HInfo"+"/"+"Drive"+"/"+"Drive-2"+"/"+"Partition"))

   println(finalPartitionPrune2)
Solution

  • This is the simplest solution I could think of:

    val finalJson = Json.toJson(
  (jsonStr \ "Start" \ "HInfo" \ "Drive")
    .as[Seq[JsValue]]
    .map(jsValue => JsObject(Seq(
      "Partition" -> (jsValue \ "Partition").get,
      "relation_tree" -> JsString(s"Start/HInfo/Drive/Drive-${(jsValue \ "Index").get}/Partition")))))

    Basically it reads all drives as sequence of JsValues and then maps them to JsObjects with needed format. It uses Index value of drive to create relation_tree value, so it will fail if this value is missing. As an alternative you can use zipWithIndex method to add your own indices to sequence. As a final step it converts sequence back to JsValue

    Here's zipWithIndex version:

    val finalJson =   Json.toJson(
  (jsonStr \ "Start" \ "HInfo" \ "Drive")
    .as[Seq[JsValue]]
    .zipWithIndex
    .map{ case (jsValue, index) => JsObject(Seq(
      "Partition" -> (jsValue \ "Partition").get,
      "relation_tree" -> JsString(s"Start/HInfo/Drive/Drive-$index/Partition")))
    })