January 2019
Beginner to intermediate
154 pages
4h 31m
English
Encoders are coded as traits in Spark 2.0. They can be thought of as an efficient means of serialization/deserialization for Spark SQL 2.0, similar to SerDes in Hive:
trait Encoder[T] extends Serializable { def schema: StructType def clsTag: ClassTag[T]}
Encoders internally convert type T to Spark SQL's InternalRow type, which is the binary row representation.
Read now
Unlock full access