cloud-fan commented on a change in pull request #23275: [SPARK-26323][SQL]
Scala UDF should still check input types even if some inputs are of type Any
URL: https://github.com/apache/spark/pull/23275#discussion_r240451835
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/expressions/UserDefinedFunction.scala
##########
@@ -88,68 +88,49 @@ sealed trait UserDefinedFunction {
private[sql] case class SparkUserDefinedFunction(
f: AnyRef,
dataType: DataType,
- inputTypes: Option[Seq[DataType]],
- nullableTypes: Option[Seq[Boolean]],
+ inputSchemas: Seq[Option[ScalaReflection.Schema]],
name: Option[String] = None,
nullable: Boolean = true,
deterministic: Boolean = true) extends UserDefinedFunction {
@scala.annotation.varargs
- override def apply(exprs: Column*): Column = {
- // TODO: make sure this class is only instantiated through
`SparkUserDefinedFunction.create()`
- // and `nullableTypes` is always set.
- if (inputTypes.isDefined) {
- assert(inputTypes.get.length == nullableTypes.get.length)
- }
-
- val inputsNullSafe = nullableTypes.getOrElse {
- ScalaReflection.getParameterTypeNullability(f)
- }
+ override def apply(cols: Column*): Column = {
+ Column(createScalaUDF(cols.map(_.expr)))
+ }
- Column(ScalaUDF(
+ private[sql] def createScalaUDF(exprs: Seq[Expression]): ScalaUDF = {
+ // It's possible that some of the inputs don't have a specific type(e.g.
`Any`), skip type
+ // check and null check for them.
+ val inputTypes = inputSchemas.map(_.map(_.dataType).getOrElse(AnyDataType))
+ val inputsNullSafe = inputSchemas.map(_.map(_.nullable).getOrElse(true))
Review comment:
Here `getOrElse` maybe better, as it matches the previous line.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
