[SQL] SPARK-1367 Fix NPE when joining parquet tables by serializing operator output.

[marmbrus]

No description provided.

[AmplabJenkins]

Merged build triggered.

[AmplabJenkins]

Merged build started.

[AmplabJenkins]

Merged build finished.

[AmplabJenkins]

Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13649/

[marmbrus]

Jenkins, test this please.

[AmplabJenkins]

Merged build triggered.

[AmplabJenkins]

Merged build started.

[AmplabJenkins]

Merged build finished. All automated tests passed.

[AmplabJenkins]

All automated tests passed.
Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/13654/

[jerryshao]

Hey Michael, it looks cool. One simple question: why HiveTableScan equivalent parameters do not specify as transient?

[marmbrus]

@jerryshao, good question. It is likely that we could mark quite a few things as transient that are currently not. However, I'm hesitant to do so this close to the 1.0 freeze since, as this bug shows, it is not always easy to reason about what is going to be required at runtime.

At some point before 1.1 we should audit all of the stuff that is being shipped to try and minimize our closure size.

[marmbrus]

I'm going to close this. @AndreSchumacher fixed it also (and actually checks to make sure the answer is correct!) in #195 and I don't want to cause conflicts.