Add example code for Read/Write Splitting sample

[congoamz]

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[brunos-bq]

do you need a new folder or a new package inside AWSDriverExample ?

[congoamz]

Sorry, I'm not sure what you are suggesting, can you clarify? I believe the ReadWriteSplittingSample directory matches our other example directory structures, eg HikariExample. Are you suggesting I move this to the AWSDriverExample directory instead of having it in its own directory? My concerns with doing that are that AWSDriverExample contains single isolated java files, but this sample might end up containing multiple java files, and the logging config in ReadWriteSplittingExample is pretty specific to just ReadWriteSplittingSample.java

[brunos-bq]

This is something that is not directly for this PR: i kinda question the structure with folders for each set of examples, instead of packages. when having multiple folders, then there are more files needed to set up the gradle configuration for each of those folders.

As this structure was in place before i added telemetry into it, i was wondering whether that was on purpose or needed to be kept. I'm ok with either, just asking questions on what makes more sense

[brunos-bq]

i think we should abandon the approach idea and have one sample file for each approach

[congoamz]

That would simplify the file, but there would also be more code duplication and users would no longer be able to view all the code at once in a single file. Do we still want to go ahead with splitting this into different files?

[brunos-bq]

Multiple files follow some arbitrary guidelines that I have seen since working on the wrapper. We tend to have multiple small files that handle one single case so users can run/copy them. If you believe the "approach" strategy matches better the way people will read it, that's completely fine by me too.

[karenc-bq]

I didnt try it out but I think you can do the following so you only execute once:

Suggested change

	try (PreparedStatement stmt = conn.prepareStatement(selectSQL)) {
	for (int i = 0; i < NUM_READS; i++) {
	stmt.setInt(1, tableNum);
	stmt.execute();
	}
	}
	try (PreparedStatement stmt = conn.prepareStatement(selectSQL)) {
	for (int i = 0; i < NUM_READS; i++) {
	stmt.setInt(1, tableNum);
	stmt.addToBatch();
	}
	stmt.executeBatch();
	}

[congoamz]

What's the advantage to doing it this way? If its just for performance, I wonder if we should make the change or not. I was using this code to purposely overload the writer so that we could observe how r/w splitting improved performance for this scenario. Then again, users probably won't be overloading their writer when they execute this code, since it requires a pretty specific configuration of settings in this file and rds cluster configuration settings. What do you think, should we go ahead with your suggested change?

[karenc-bq]

It is a more common practice to use batch execute because of performance improvements. I just think that since it is a sample code we would want to follow best practices in our code. If we use batchExecute do we no longer see a performance improvement?