Github user kiszk commented on a diff in the pull request:
https://github.com/apache/spark/pull/17302#discussion_r106916774
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/ExistingRDD.scala ---
@@ -70,7 +70,20 @@ object RDDConversions {
object ExternalRDD {
def apply[T: Encoder](rdd: RDD[T], session: SparkSession): LogicalPlan =
{
- val externalRdd = ExternalRDD(CatalystSerde.generateObjAttr[T],
rdd)(session)
+ val attr = {
+ val attr = CatalystSerde.generateObjAttr[T]
--- End diff --
When `ExternalRDD.apply` calls `CatalystSerde.generateObjAttr`,
`deserializer.nullable` has not been resolved yet. In other call path to
`CatalystSerde.generateObjAttr`, `deserializer.nullable` may be already
resolved.
I think that this problem depends on the calling context. If we fix
`CatalystSerde.generateObjAttr` by using a general approach, the following
approach is possible:
1. call `deserializer.nullable` in `CatalystSerde.generateObjAttr`
2. If an exception does not occurs, use a value of `deserializer.nullable`
3. If an unresolved exception occurs, infer `nullable` based on
`deserializer.dataType`
Is this OK with you instead of modifying `ExternalRDD.apply()`?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
