hector scala type mismatch
What could be wrong?
val is = IntegerSerializer.get
mutator.addInsertion(deviceId, COLUMN_FAMILY_CARSTATUS, createColumn("mileage", 111, ss, is))}
ModelOperation.scala:96: error: type mismatch;
[INFO] found : me.prettyprint.cassandra.serializers.IntegerSerializer
[INFO] required: me.prettyprint.hector.api.Serializer[Any]
[INFO] Note: java.lang.Integer <: Any (and me.prettyprint.cassandra.serializers.IntegerSerializer <: me.prettyprint.cassandra.serializers.AbstractSerializer[java.lang.Integer]), but Java-defined trait Serializer is invariant in type T.
[INFO] You may wish to investigate a wildcard type such as `_ <: Any`. (SLS 3.2.10)
[INFO] mutator.addInsertion(deviceId, COLUMN_FAMILY_CARSTATUS, 开发者_JAVA百科createColumn("mileage", 111, ss, is))}
The error is saying that createColumn
requires a serializer of type Serializer[Any]
, but you're passing one of type Serializer[Integer]
. This would only work if Serializer
were covariant in its type parameter (i.e., defined as Serializer[+T]
). But instead, Serializer
comes from Java, where covariance works differently.
The type Serializer[Integer]
can be safely cast to Serializer[_ <: Any]
, so the Scala compiler is suggesting that maybe createColumn
should have been written to expect that less specific wildcard type instead.
If you can't modify createColumn
, then a last resort is to use the "type system escape hatch" asInstanceOf
to cast to the expected type:
val is = IntegerSerializer.get.asInstanceOf[Serializer[Any]] // defeats type system
mutator.addInsertion(... is ...)
精彩评论