Consider changing tabular formatting default for data.table's `fcase()`

[gshelor]

If you wouldn't do it for dplyr's case_when(), you shouldn't be doing it for its comparable functions in other packages

[lionel-]

I wouldn't say fcase is comparable to case_when, syntactically speaking. The former is a two-colum tabular arrangement of arguments, whereas the latter is only a one-column arrangement (the formulas count as single arguments).

You can remove the default handling for fcase in your toml file:

default-table = false

[gshelor]

Maybe, but that's not a meaningful distinction when it comes to the readability of the code, and their functionality is so similar that formatting them differently makes no sense anyway.

[lionel-]

I respectfully disagree. And ideally we might even want to format the case_when() calls with alignment at the ~, LHS right-aligned, and RHS left-aligned. So a quite different formatting.

That's not to say our current handling of fcase() is the best choice, but I think we'd need more stronger reasons to change the default.

[lionel-]

By the way the only viable alternative to tabular formatting is no formatting at, i.e. we'd add ftable to the set of functions that are automatically treated as fmt: skip.

Otherwise this:

fcase(
  x < 5L, 1L,
  x > 5L, 3L
)

would get expanded to:

fcase(
  x < 5L,
  1L,
  x > 5L,
  3L
)