Return a tibble instead of a data.frame

[wibeasley]

For a while, REDCapR functions has been using readr::read_csv() for a while, but has continued to (upcast to) return data.frame. There's a small chance that someone's code (that pulls a single column from a REDCap project) would break because they depend on something like ds[, 3] returning a vector. That's the behavior of a data.frame. In contrast for tibbles, ds[, 3] returns a single-column rectangle/tibble/data.frame.

Until now, I've felt the tradeoff of (the small chance of breaking backward compatibility) favored returning a data.frame. I've changed my mind the past few weeks b/c calling readr::problems(ds) is informative if ds is a tibble, but not if ds is a data.frame.

I want those parsing problems to be as easy as possible for users to determine. Returning a tibble (instead of a data.frame)