从 Scala 中的 Dataframe 的嵌套结构数组中选择几列

Question

I have a dataframe with array of struct and inside that another array of struct. Any easy way to select few of the structs in the main array and also few in the nested array without disturbing the structure of the entire dataframe?

SIMPLE INPUT:

-MainArray
---StructCol1
---StructCol2
---StructCol3
---SubArray
------SubArrayStruct4
------SubArrayStruct5
------SubArrayStruct6

SIMPLE OUTPUT:

-MainArray
---StructCol1
---StructCol2
---SubArray
------SubArrayStruct4
------SubArrayStruct5

The source code to try it is as below

import org.apache.spark.sql.types.StructField
import org.apache.spark.sql.types.StructType
import org.apache.spark.sql.types.StringType
import org.apache.spark.sql.types.ArrayType
import org.apache.spark.sql.types.IntegerType

val arrayStructData = Seq(
      Row("Army",List(Row("1","Infantry","100",List(Row("Gun","Station"),Row("Bazooka","Barracks"))),Row("2","Cavalry","150",List(Row("Grenadier","Seige factory"),Row("Canon","Tank Factory"))))),
      Row("Navy",List(Row("3","Transport","200",List(Row("Cruiser","Cruise Lines"),Row("SubMarine","Yard"))),Row("4","Battle Ships","250",List(Row("Frigate","Dock"),Row("Galleon","Hub")))))
    )


val arrayStructSchema = new StructType()
      .add("Category",StringType)
      .add("ArmyOrNavy",ArrayType(new StructType()
        .add("ID",StringType)
        .add("Type",StringType)
        .add("Count",StringType)
        .add("Items",ArrayType(new StructType().add("ItemName",StringType).add("ItemTrainingArea",StringType)))
        ))


val df = spark.createDataFrame(spark.sparkContext.parallelize(arrayStructData),arrayStructSchema)
    
df.printSchema()
df.show(false)

    root
 |-- Category: string (nullable = true)
 |-- ArmyOrNavy: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- ID: string (nullable = true)
 |    |    |-- Type: string (nullable = true)
 |    |    |-- Count: string (nullable = true)
 |    |    |-- Items: array (nullable = true)
 |    |    |    |-- element: struct (containsNull = true)
 |    |    |    |    |-- ItemName: string (nullable = true)
 |    |    |    |    |-- ItemTrainingArea: string (nullable = true)

+--------+-----------------------------------------------------------------------------------------------------------------------------------+
|Category|ArmyOrNavy                                                                                                                         |
+--------+-----------------------------------------------------------------------------------------------------------------------------------+
|Army    |[[1, Infantry, 100, [[Gun, Station], [Bazooka, Barracks]]], [2, Cavalry, 150, [[Grenadier, Seige factory], [Canon, Tank Factory]]]]|
|Navy    |[[3, Transport, 200, [[Cruiser, Cruise Lines], [SubMarine, Yard]]], [4, Battle Ships, 250, [[Frigate, Dock], [Galleon, Hub]]]]     |
+--------+-----------------------------------------------------------------------------------------------------------------------------------+

The output I need is

    root
 |-- Category: string (nullable = true)
 |-- ArmyOrNavy: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- ID: string (nullable = true)
 |    |    |-- Items: array (nullable = true)
 |    |    |    |-- element: struct (containsNull = true)
 |    |    |    |    |-- ItemTrainingArea: string (nullable = true)

I tried doing something like this but this doesn't look right

val df2 = df.selectExpr("Category",
  "Array (Struct(ArmyOrNavy.ID,CAST(ArmyOrNavy.Items AS array<array<struct<ItemName:string,ItemTrainingArea:string>>>) Items))  as ArmyOrNavy")
df2.printSchema
df2.show(false)

Answer 1

You can do it using to_json and from_json and set new struct DateType for struct field ( array ) while parsing json:

val newArrayType = ArrayType(
  new StructType()
    .add("ID", StringType)
    .add("Items", ArrayType(
      new StructType()
        .add("ItemTrainingArea", StringType)
    ))
)
val jsonFieldName = "ArmyOrNavy_json"
val transformedDF = df.withColumn(jsonFieldName, to_json($"ArmyOrNavy"))
  .withColumn("ArmyOrNavy", from_json(col(jsonFieldName), newArrayType))
  .drop(jsonFieldName)
transformedDF.printSchema()
transformedDF.show(truncate = false)


// output
root
 |-- Category: string (nullable = true)
 |-- ArmyOrNavy: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- ID: string (nullable = true)
 |    |    |-- Items: array (nullable = true)
 |    |    |    |-- element: struct (containsNull = true)
 |    |    |    |    |-- ItemTrainingArea: string (nullable = true)
+--------+----------------------------------------------------------------------+
|Category|ArmyOrNavy                                                            |
+--------+----------------------------------------------------------------------+
|Army    |[[1, [[Station], [Barracks]]], [2, [[Seige factory], [Tank Factory]]]]|
|Navy    |[[3, [[Cruise Lines], [Yard]]], [4, [[Dock], [Hub]]]]                 |
+--------+----------------------------------------------------------------------+