Updated testing for instatiation of an epi_archive

[kenmawer]

No description provided.

[brookslogan]

Hopefully just one more test to add (above). Please also address the warnings emitted by check() when running the tests:
https://github.com/cmu-delphi/epiprocess/pull/159/checks#step:5:245

We should either capture/muffle these warnings, or pass/set things up so they are never generated (e.g., by passing compactify=TRUE if these are all just compactification warnings)

[brookslogan]

Triggering GitHub to mark my review as complete, so that you can re-request review once again when you think these final couple of items are complete. (Trying to keep on top of reviews using https://github.com/pulls/review-requested)

[brookslogan]

Also, I realize that the as_epi_archive behavior here on a keyed DT when we don't pass other_keys seems a bit fishy. If they've keyed the thing appropriately with geo_type, time_type, {other keys}, version already, I believe this means that we drop all their {other keys} despite the data table key looking like a valid epi_archive DT key. Probably we want some type of warning here at least. Could be a separate issue.

[kenmawer]

I agree with the most recent comment. However, I believe I have already seen this as an issue.

[kenmawer]

Note that I may be mixing up a different issue with what should be addressed with the key being dropped.

[brookslogan]

I didn't check to see if this was already filed as an Issue or not. Could you please confirm whether or not it is already filed?

[brookslogan]

Also I realize I never "unresolved" the conversation with the "one more test to add (above)" so it was visible again --- please take a look.

[kenmawer]

I didn't see the issue elsewhere about keys not being retained, so I filed it as #201.