I'm trying to call the functions.typedLit from Spark library in my other question. And it asks for two parameters, a literal object and its type (a TypeTag).
The first argument, I can come up with myself (I think). But as for the second one, I don't know how to instantiate an object for it. I'm trying to pass the TypeTag> as the second parameter.
Can someone please tell how to create the TypeTag for Seq in Java? Here's the code I'm trying to complete:
Seq seq = JavaConverters
.asScalaIteratorConverter((new ArrayList()).iterator())
.asScala()
.toSeq();
Column litColumn = functions.>typedLit(seq, ???)
解决方案
In your other question I said
and supplying the CanBuildFrom argument from Java is... technically possible, but not something you want to do.
Unfortunately, for TypeTag it's even less possible: unlike CanBuildFrom, they aren't supplied by a library, but built into Scala compiler.
The best advice I can give is to create a Scala file supplying the type tags you need to use from Java, since you only should need a limited number of them:
object TypeTags {
val SeqInteger = typeTag[Seq[Integer]]
...
// or
val Integer = typeTag[Integer]
def Seq[A](implicit tt: TypeTag[A]) = typeTag[Seq[A]]
}
and then from Java TypeTags.SeqInteger or TypeTags.Seq(TypeTags.Integer).
In other places Spark provides special API for use from Java (look for .java packages), but I couldn't find one for functions.typedLit.