Github user hvanhovell commented on a diff in the pull request:
https://github.com/apache/spark/pull/20756#discussion_r172801767
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/objects/objects.scala
---
@@ -1261,8 +1261,24 @@ case class InitializeJavaBean(beanInstance:
Expression, setters: Map[String, Exp
override def children: Seq[Expression] = beanInstance +:
setters.values.toSeq
override def dataType: DataType = beanInstance.dataType
- override def eval(input: InternalRow): Any =
- throw new UnsupportedOperationException("Only code-generated
evaluation is supported.")
+ override def eval(input: InternalRow): Any = {
+ val instance = beanInstance.eval(input).asInstanceOf[Object]
+ if (instance != null) {
+ setters.foreach { case (setterMethod, fieldExpr) =>
--- End diff --
Why are we resolving setters during `eval`? That seems a bit expensive. How
about we create the setters before we execute eval? For example (I got a bit
carried away):
```scala
private lazy val resolvedSetters = {
val ObjectType(beanClass) = beanInstance.dataType
val lookup = MethodHandles.lookup()
setters.map {
case (name, expr) =>
// Resolve expression type (should be better!)
val fieldClass =
CallMethodViaReflection.typeMapping(expr.dataType).head
val handle = lookup.findVirtual(
beanClass,
name,
MethodType.methodType(classOf[Unit], fieldClass))
handle -> expr
}
}
override def eval(input: InternalRow): Any = {
val bean = beanInstance.eval(input)
resolvedSetters.foreach {
case (setter, expr) =>
setter.invoke(bean, expr.eval(input))
}
bean
}
```
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
