NA is a boolean?

[richelbilderbeek]

Boy, did I enjoy finding this package! I felt stupid when writing my own version of is_a_bool (https://github.com/ropensci/beautier/blob/master/R/is_one_bool.R#L15) and was relieved to call assertive instead of maintaining my handcrafted functions.

A showstopper to do so, however, is this:

assertive::assert_is_a_bool(NA)

I expect a boolean to be either TRUE or FALSE; NA is not among the set of valid booleans. Yet, I see in the documentation that classifying an NA as a Boolean appears to be intentional.

Is it the intention to classify NA as a boolean?

• If yes, I hope the reasoning for this will be added to the documentation. I will -sadly- have to use my own version of is_a_bool :-(
• If no, I hope NA will be classified as not being a boolean and I can look forward to use assertive’s next version

Whatever the case, I am happy to see this package exists. Keep up the good work

[Richard Cotton]

@richelbilderbeek

This is the intended behavior since NA is logical. You can see this using class(NA).

After creating the function I realized that it's a fairly common thing to want to check for a value being TRUE or FALSE in if conditions, so assertive.code also has is_if_condition() which does what you want. See

https://www.rdocumentation.org/packages/assertive.code/topics/assert_is_if_condition

[Richard Cotton]

The behavior is intentional, and there is an alternative function, so closing this.

[richelbilderbeek]

It appears I underestimated the awesomeness of your work