DataType (UDT) v.s. Encoder in Spark SQL

Question

In Spark SQL, there\'re limited DataTypes for Schema, and there\'re limited Encoders for converting JVM objects to and from the internal Spark SQL representation.