NEW DataQuery::exists now generates EXISTS SQL statements

[ScopeyNZ]

This PR adds an exists method to DataQuery that will generate an EXISTS statement from the existing query. This basically looks like this:

SELECT EXISTS(SELECT * FROM MyDataObject WHERE SomeProperty)

This is the preferred way to do these sorts of queries as the engine can just do checks like "does the MyDataObject.SomeProperty index have any entries for true" - this can be a lot faster than counting results.

[ScopeyNZ]

I just realised I forgot to ensure tests are passing 🤦‍♂️ . I'll hope travis has good news.

[ScopeyNZ]

Ok the remaining failing tests are fixed by @sminnee's recent updates to consistent boolean return types and relies on a fix in 2.x of the postgres module.

Should I update travis config or update my code to be compatible with older versions of postgres?

[robbieaverill]

Should I update travis config or update my code to be compatible with older versions of postgres?

It sounds like that should be done upstream, but you could patch it here too if you want to

[kinglozzer]

Looks like a great addition! If it’s possible to work around the BC issue I’d be happy to take this in

[ScopeyNZ]

Ok I've updated this PR to only contain the EXISTS query changes. I'll raise additional PRs for the threshold fix and the idea about the "prefer conditions" API.

[ScopeyNZ]

Hmmmn.. Tests...

I have no idea why it's failing. The 5.6 build is pretty much the same (runs on postgres) and it's passing...

[ScopeyNZ]

I figured out the failing tests: #8934

[ScopeyNZ]

I added a very simple test. I tried doing some craziness with aggregates and HAVING but the API actually makes it really hard to do as you can't really select the field you're aggregating in your HAVING clause and that's not so easy to with the DataList API.

[emteknetnz]

This pull request hasn't had any activity for a while. Are you going to be doing further work on it, or would you prefer to close it now?

[robbieaverill]

@emteknetnz this isn't waiting for the author

[sminnee]

I resolved the conflict but I'd probably give such an old branch a rebase before merging.

[sminnee]

...rebased against 4 and updated the merge target.

[sminnee]

If the tests pass I'm happy with this.