HeartSaVioR commented on a change in pull request #23854: [SPARK-22000][SQL]
Address missing Upcast in JavaTypeInference.deserializerFor
URL: https://github.com/apache/spark/pull/23854#discussion_r259650224
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/JavaTypeInference.scala
##########
@@ -194,13 +195,31 @@ object JavaTypeInference {
*/
def deserializerFor(beanClass: Class[_]): Expression = {
val typeToken = TypeToken.of(beanClass)
- deserializerFor(typeToken, GetColumnByOrdinal(0,
inferDataType(typeToken)._1))
+ val walkedTypePath = s"""- root class: "${beanClass.getCanonicalName}""""
:: Nil
+ val (dataType, nullable) = inferDataType(typeToken)
+
+ // Assumes we are deserializing the first column of a row.
+ val input = upCastToExpectedType(
+ GetColumnByOrdinal(0, dataType), dataType, walkedTypePath)
+
+ val expr = deserializerFor(typeToken, input, walkedTypePath)
+ if (nullable) {
+ expr
+ } else {
+ AssertNotNull(expr, walkedTypePath)
+ }
Review comment:
Yes that sounds better, wondering which name of object (maybe) is correct.
Would `DeserializerHelper` make sense? I'll try to refactor.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
