[SPARK-4774] [SQL] Makes HiveFromSpark more portable

[ksakellis]

HiveFromSpark read the kv1.txt file from SPARK_HOME/examples/src/main/resources/kv1.txt which assumed
you had a source tree checked out. Now we copy the kv1.txt file to a temporary file and delete it when
the jvm shuts down. This allows us to run this example outside of a spark source tree.

[vanzin]

I don't think this works for non-empty directories. You'd need to recursively delete it. Instead of creating a temp directory, I'd just use File.createTempFile(), unless the name of the file reeeeeally needs to be kv1.txt.

Also, the paranoid in me would close all open streams in a finally block, but not super important in this code.

[vanzin]

Looks ok aside from the cleanup issue.

[SparkQA]

Test build #24201 has started for PR 3628 at commit fca97ee.

• This patch merges cleanly.

[marmbrus]

Thanks for doing this! Mind adding [SQL] to the PR title?

[SparkQA]

Test build #24202 has started for PR 3628 at commit 6770f83.

• This patch merges cleanly.

[SparkQA]

Test build #24201 has finished for PR 3628 at commit fca97ee.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24201/
Test PASSed.

[SparkQA]

Test build #24202 has finished for PR 3628 at commit 6770f83.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24202/
Test PASSed.

[marmbrus]

LGTM