Skip to content

Latest commit

 

History

History
144 lines (113 loc) · 6.14 KB

6_advanced_mapper.md

File metadata and controls

144 lines (113 loc) · 6.14 KB
Raw

Documentation

Customizing the mapping between Scala and Cassandra

While mapping rows to tuples and case classes work out-of-the box, in some cases you might need more control over Cassandra-Scala mapping. For example Java classes are likely to use the JavaBeans naming convention, where accessors are named with get, is or set prefixes.

To customize column-property mappings, you need to put an appropriate ColumnMapper[YourClass] implicit object in scope. Define such an object in a companion object of the class being mapped. A ColumnMapper affects both loading and saving data. A few special ColumnMapper implementations are included.

Working with JavaBeans

To work with Java classes, use JavaBeanColumnMapper. Make sure your objects are Serializable, otherwise Spark won't be able to send them over the network.

import com.datastax.spark.connector.mapper.JavaBeanColumnMapper
class WordCount extends Serializable { 
    private var _word: String = ""
    private var _count: Int = 0
    def setWord(word: String) { _word = word }
    def setCount(count: Int) { _count = count }
    override def toString = _word + ":" + _count
}

object WordCount {
    implicit object Mapper extends JavaBeanColumnMapper[WordCount] 
}

sc.cassandraTable[WordCount]("test", "words").toArray
// Array(bar:20, foo:10)

To save objects of class WordCount, you'll need to define getters.

Manually specifying the property-name to column-name relationship

If for some reason you wish to associate a column of a different name than the property, you may pass a column translation Map to a DefaultColumnMapper or JavaBeanColumnMapper:

case class WordCount(w: String, c: Int)

object WordCount { 
    implicit object Mapper extends DefaultColumnMapper[WordCount](
        Map("w" -> "word", "c" -> "count")) 
}

sc.cassandraTable[WordCount]("test", "words").toArray
// Array(WordCount(bar,20), WordCount(foo,10))

sc.parallelize(Seq(WordCount("baz", 30), WordCount("foobar", 40)))
  .saveToCassandra("test", "words", SomeColumns("word", "count"))

Writing custom ColumnMapper implementations

To define column mappings for your classes, create an appropriate implicit object implementing ColumnMapper[YourClass] trait. This API is subject to future changes, so please refer to the current ScalaDoc.

Using custom field types

To map a Cassandra column to a field of user-defined type, register custom TypeConverter implementations. For example, imagine you want emails to be stored in custom Email class, wrapping a string:

case class EMail(email: String)

To tell the connector how to read and write fields of type EMail, you need to define two type converters - from String to Email and from Email to String:

import com.datastax.spark.connector.types._
import scala.reflect.runtime.universe._

object StringToEMailConverter extends TypeConverter[EMail] {
  def targetTypeTag = typeTag[EMail]
  def convertPF = { case str: String => EMail(str) }
}

object EMailToStringConverter extends TypeConverter[String] {
  def targetTypeTag = typeTag[String]
  def convertPF = { case EMail(str) => str }
}
    
TypeConverter.registerConverter(StringToEMailConverter)
TypeConverter.registerConverter(EMailToStringConverter)

Now you can map any Cassandra text or ascii column to EMail instance. The registration step must be performed before creating any RDDs you wish to use the new converter for.

Additionally, defining the StringToEMailConverter as an implicit object allows to use generic CassandraRow#get with your custom EMail field type.

The following table specifies the relationship between a Cassandra column type and the type of needed TypeConverter.

Cassandra column type | Object type to convert from / to

ascii | java.lang.String
bigint | java.lang.Long
blob | java.nio.ByteBuffer boolean | java.lang.Boolean
counter | java.lang.Long
decimal | java.math.BigDecimal double | java.lang.Double
float | java.lang.Float
inet | java.net.InetAddress int | java.lang.Integer
text | java.lang.String timestamp | java.util.Date uuid | java.util.UUID timeuuid | java.util.UUID varchar | java.lang.String varint | java.math.BigInteger user defined | com.datastax.spark.connector.UDTValue

Custom converters for collections are not supported.

When defining your own TypeConverter make sure it is Serializable and works properly after being deserialized. For example, if you want to register a custom TypeConverter producing Cassandra UDT values, you must register a converter converting to connector's UDTValue class, not Java Driver's UDTValue. This is because many Java Driver's classes are not Serializable and would not be possible for such converter to be properly serialized/deserialized.

Low-level control over mapping

The ColumnMapper API cannot be used to express every possible mapping, e.g., for classes that do not expose separate accessors for reading/writing every column.

For converting a low-level Row object obtained from the Cassandra Java driver into an object stored in RDD, this Spark driver uses a RowReader instance. An appropriate RowReader is obtained from an implicit RowReaderFactory resolved based on the target RDD item type. You need to provide a custom implicit RowReaderFactory and RowReader for working with your class, and have it in scope when calling cassandraTable.

In the same way, when writing an RDD back to Cassandra, an appropriate implicit RowWriterFactory and RowWriter are used to extract column values from every RDD item and bind them to an INSERT PreparedStatement.

Please refer to the ScalaDoc for more details.

Next - Using Connector in Java