Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/6361#issuecomment-109171615
Trying to dig into the stracktrace a bit:, here an excerpt from
GenericUDTFExplode.java in hive-exec-0.13.1a-sources.jar:
```java
private transient final Object[] forwardListObj = new Object[1];
private transient final Object[] forwardMapObj = new Object[2];
@Override
public void process(Object[] o) throws HiveException {
switch (inputOI.getCategory()) {
case LIST:
ListObjectInspector listOI = (ListObjectInspector)inputOI;
List<?> list = listOI.getList(o[0]);
if (list == null) {
return;
}
for (Object r : list) {
forwardListObj[0] = r; // <------- line 93 that throws NPE
forward(forwardListObj);
}
break;
```
Based on the line number, I think that the NPE is being thrown because the
transient field `forwardListObj` is null`. Since this field is `transient
final` I wonder if this class wasn't designed to have its `process` method
called on deserialized instances. If that's the case, though, I'm puzzled at
why we wouldn't see similar behavior with Java serialization. The problem must
be a bit deeper and merits more investigation.
---
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]
