port jdbc examples to hsqldb

[sammefford]

We created four jdbc cookbook examples: BulkLoadFromJdbcRaw, BulkLoadFromJdbcWithJoins, BulkLoadFromJdbcWithSimpleJoins, and IncrementalLoadFromJdbc. We used mysql (mariadb) to present a real-world scenario of migrating data between relational and MarkLogic.

We also created unit test wrappers for them so we can ensure the cookbook examples work against the latest client & server.: BulkLoadFromJdbcRawTest, BulkLoadFromJdbcWithJoinsTest, BulkLoadFromJdbcWithSimpleJoinsTest, and IncrementalLoadFromJdbcTest. However, that would require our Jenkins and regression environments to install mysql. Rather than do that, let's try hsqldb since it can run embedded in the JVM. Ideally, we'd run the tests off some data in src/main/resources/. I'd like the code to work unmodified against mysql and hsqldb if possible.

[grechaw]

+1 I like this plan

[grechaw]

If there were a way to coordinate this work with ml-gradle, then you could deploy, run, and distribute code for a cookbook example as its own project, rather than embedded in the Client API repository. That might be something to consider when a larger refactoring opportunity arises.

[sammefford]

I could see a github project containing just cookbook examples. But would it make sense to have a separate project for each cookbook example?

Also, does ml-gradle help with managing java code? I thought it was for MarkLogic. Or did you mean gradle?

[jmakeig]

But would it make sense to have a separate project for each cookbook example?

Probably not. Separate repo for all of the recipes is probably where I’d start.

Also, does ml-gradle help with managing java code?

No, but gradle’s pretty good at that.

[grechaw]

ml-gradle helps if you have extensions (REST transforms, service extensions) to install.