Remove reference to data.table from tests

[jwbowers]

Proposing this patch as a response to an email from Kurt H:

These seem to have undeclared package dependencies in their unit test code (R files in tests subdirs), see below.

Can you pls fix as necessary? (Add the missing package dependencies to Suggests, I guess.)

Please note that these issues are currently not yet detected by the CRAN incoming (or regular) checks.

Best -k

$RItools
'library' or 'require' call not declared from: ‘data.table’

[benthestatistician]

1. This is an annoying thing to have to deal with. Thanks for taking initiative on addressing it, Jake.
2. I'd much prefer a solution that preserves the future-proofing we get from the test @josherrickson added in 273b6da .
3. To that end, I'd propose instead to set up a special test file, "tests/testthat/test.notforCRAN.R" or similar, where we collect all the tests we're not ready to get rid of ourselves but that we don't want to give to CRAN. If we add it to ".Rbuildignore", CRAN should never know about it.

[benthestatistician]

I've pushed a new branch w/ my preferred solution, creatively named jpbowers-data.table-patch-2. I'm going to close out this merge request and start a new one. Of course this one (or perhaps a third new one) can be reopened as needed.