This is a section about managing custom type support in phantom. It coveralls modelling scenarios where you wish to make your types act as primitive to Cassandra. There are several options at hand.
JSON Columns
One simple way to encode case classes or other Scala types as Cassandra native is to use native JSON support in phantom. This works in a really simple way, phantom will automatically use implicits to serialize/de-serialize your types to something Cassandra understands, namely string types.
Let’s explore a simple example using the circe library. Phantom does not ship with any particular JSON library, you have complete freedom over what JSON library you use.
import com.outworkers.phantom.dsl._
import io.circe._
import io.circe.generic.semiauto._
import io.circe.parser._
import io.circe.syntax._
case class JsonRecord(
prop1: String,
prop2: String
)
object JsonRecord {
implicit val jsonDecoder: Decoder[JsonRecord] = deriveDecoder[JsonRecord]
implicit val jsonEncoder: Encoder[JsonRecord] = deriveEncoder[JsonRecord]
implicit val jsonPrimitive: Primitive[JsonRecord] = {
Primitive.json[JsonRecord](_.asJson.noSpaces)(decode[JsonRecord](_).right.get)
}
}
case class JsonClass(
id: UUID,
name: String,
json: JsonRecord,
optionalJson : Option[JsonRecord],
jsonList: List[JsonRecord],
jsonSet: Set[JsonRecord]
)
abstract class JsonTable extends Table[JsonTable, JsonClass] {
object id extends UUIDColumn with PartitionKey
object name extends StringColumn
object json extends JsonColumn[JsonRecord]
object optionalJson extends OptionalJsonColumn[JsonRecord]
object jsonList extends JsonListColumn[JsonRecord]
object jsonSet extends JsonSetColumn[JsonRecord]
}