[SPARK-16303][DOCS][EXAMPLES][WIP] Updated SQL programming guide and examples

[aokolnychyi]

What changes were proposed in this pull request?

• Hard-coded Spark SQL sample snippets were moved into source files under examples sub-project.
• Removed the inconsistency between Scala and Java Spark SQL examples
• Scala and Java Spark SQL examples were updated

How was this patch tested?

The work is still in progress. All involved examples were tested manually. An additional round of testing will be done after the code review.

[aokolnychyi]

@liancheng could you, please, review this PR?

[aokolnychyi]

Is it the correct way to refresh?

[liancheng]

Yes.

[liancheng]

test this please

[liancheng]

add to whitelist

[SparkQA]

Test build #62092 has finished for PR 14119 at commit 95f0f41.

• This patch fails RAT tests.
• This patch merges cleanly.
• This patch adds no public classes.

[liancheng]

Since you've added JavaSparkSqlExample.scala, we can remove JavaSparkSQL.scala now. (I guess that file was from my original WIP branch?)

[liancheng]

Please add Apache license header.

[liancheng]

This looks pretty good! Only found a few minor issues. Thanks for working on it!

[liancheng]

Can we add actual stdout output after each .show() call?

[aokolnychyi]

Thanks for the review, I really appreciate. I tried to take into account all comments and update the PR accordingly.

Summary of the updates

• JavaSparkSQL.java file was removed. I kept it initially since the file itself was quite old (2+ years) and it was present in your original WIP branch alongside the new file. But I can confirm that the new file covers the same functionality and more. No need to keep the old one, agree with you.
• Apache header in JavaSqlDataSourceExample.java was added.
• $-notation instead of df("...") in Scala examples.
• col("...") instead of df.col("...") in Java examples.
• Blank lines before {% include_example programmatic_schema ... } were added. However, everything was rendered fine locally even without them.
• 2 space indentation for chained method calls. My fault, sorry.
• Actual outputs for all show() calls were added.
• Tested manually and via ./dev/run-tests.

Open questions

• Shall I add blank lines before each {% include_example ... } or only before those two examples?
• I pointed to a wrong location that exceeded the length limit. It is exactly the same functionality but in Java. So, 113 and 117 lines of the JavaSqlDataSourceExample.java file. In my view, it would make sense to keep them as they are now for the better looking documentation.

[SparkQA]

Test build #62192 has finished for PR 14119 at commit 7451fc7.

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

[liancheng]

LGTM, I've merged this to master and branch-2.0. Thanks for working on this!

I only observed one weird rendering case caused by the missing blank lines before {% include_example %}, but maybe my local Jekyll version is too low. I think it's fine to leave other lines as is. The exceeded lines in the Java example file should be OK.

Could you please remove the WIP tag from the PR title? (I've removed it manually while merging this PR.)