dealing with NAs

[moodymudskipper]

dealing with NAs

[karoliskoncevicius]

hmm to me it looks a bit too complicated.

If NAs are not desirable we can actually switch that particular behaviour to match %in% instead of ==. But then the inconsistency with [== will still remain.

Another alternative is to provide slightly different operator syntax for dealing with NAs differently. Maybe capital letters %IN{}%, like a "strict" %in{}% But not sure if it's intuitive.

[moodymudskipper]

I think I'd rather not multiply by two our whole set of operators, I miss these functions in a CRAN package, but this might not be the place for them.

[karoliskoncevicius]

Yup, agreed. What is your opinion about the other option - breaking consistency with == just for NA values?

[moodymudskipper]

I'm not sure about it, it's easy to remove NAs or transform them, if they're gone by default you won't get them back.

unrelated FYI : rstudio/rstudio#5767

[moodymudskipper]

I don't completely dislike the capital letter idea though! We can think about it as we use the package.

[karoliskoncevicius]

Yup I agree about not getting NA values back if we get rid of it. Very good point.

About rstudio issue - great catch. I do not use R-studio, so do not have this particular problem. But it would be cumbersome for potential users of the package.