Davis Vaughan

Probably same issue as https://github.com/HowardHinnant/date/issues/563

> Yes I would have expected evaluation to stop when the first condition that is true has been met That's a separate request. You're asking for _lazy_ semantics here, which...

First off, just know that I do agree with you here. My ideal version of `case_when()` would follow two rules: - The LHSs must all have the same `size`. No...

Also, in `case_match()` the behavior is more obvious. We get a primary input, `.x`, so we get to assert that all the RHSs should have size 1 or size equal...

This also seems to only really affect things when the LHSs all have length 1, causing them to get recycled to any other size (i.e. the size 0 from `character()`...

See also #7088 where RHS recycling could possibly be used in some cases to incorrectly mimic a simple if statement. I'd argue this should be an error, and changing the...

@benhmin if you are using `case_when()` for single `TRUE` / `FALSE` values, then I would _highly_ suggest you use a simple if/else instead. if/else is absolutely the way to go...

Like, the outcome possibilities are `"A"`, `"B"`, or `"C"`, but its always a single scalar value? In that case, I'd say you just want a "simple" `switch()` statement FWIW -...

There are 4 (ish) verbs that have data-masking semantics when I sort of feel they probably should have had tidy-select semantics, those are: - `distinct()` - `count()` - `arrange()` -...

> Is the idea that people want to tidy-select with these functions more often than data-mask? Yes definitely this. Making these functions data-masking (i.e. allowing pretty much any arbitrary expression)...