我有两个数据集要连接
df

|-- key: struct (nullable = false)
 |    |-- name: string (nullable = true)
 |    |-- subId: struct (nullable = true)
 |    |    |-- x: integer (nullable = false)
 |    |    |-- y: integer (nullable = false)
 |    |    |-- level: integer (nullable = false)
 |-- otherItems: struct (nullable = false)
 |    |-- nameRestaurant: string (nullable = true)
 |    |-- thing: struct (nullable = true)

和另一个df2

|-- key: struct (nullable = false)
 |    |-- name: string (nullable = true)
 |    |-- subId: struct (nullable = true)
 |    |    |-- x: integer (nullable = false)
 |    |    |-- y: integer (nullable = false)
 |    |    |-- level: integer (nullable = false)
 |-- attribute: struct (nullable = false)
 |    |-- address: string (nullable = true)
 |    |-- someThing: struct (nullable = true)

我需要在key列上连接两个数据集，即
val df3 = df.join(df2, Seq("key"), "left")
然而，执行该连接会导致没有匹配，而我确信它们存在
当我尝试通过执行以下操作扩展联接时

val df3 = df.join(df2, Seq("key.name", "key.subId.x", "key.subId.y", "key.subId.level"), "left")

我收到错误

org.apache.spark.sql.AnalysisException: USING column `key.name` cannot be resolved on the left side of the join.

在结构体下面的项上加入是不可能的吗？有人能建议最好的方法吗？