hughjonesd/santoku

Clash with chop() in tidyr

cpsyctc2 opened this issue · 6 comments

cpsyctc2 commented

I'm not sure if this is a bug but it may be a challenge if you want to get santoku very wide uptake. I think your chop is great but I've just realised that it collides with chop() in tidyR. Obviously it's easy to work around this by using santoku::chop() when I've loaded tidyR first but perhaps a name contention with such a widely used package is unwise and maybe it's not too late to change? Very best, Chris

hughjonesd commented
cpsyctc2 commented

Not on twitter but best wishes with your work.

hughjonesd commented

In the end I decided to leave it as it is, but added kut() as a synonym. chop() is very intuitive, and I can't find a nice synonym that is as memorable. chop() is a second-order, rather technical part of the latest tidyr API, and I wouldn't be surprised if it vanished. So, I will brazen it out. And if the tidyverse guys complain, I can always remind them of filter()...

Leaving open for the moment to remind me that this issue is not dead.

dpprdan commented

Was gonna suggest kiru() (+1 from me in any case) and mention the the problematic meaning of kut() in Dutch, but looked up the Twitter thread and saw that those already have been mentioned.

dpprdan commented

FWIW: While I agree with the "second-order part of the tidyr API" (not sure about the vanishing part, though), one cosmetic aspect worth considering might be that if you load both packages, you'll always get a masked warning (unless you explicitly set warn.conflicts = FALSE):

library(tidyr)
library(santoku)
#> 
#> Attaching package: 'santoku'
#> The following object is masked from 'package:tidyr':
#> 
#>     chop
hughjonesd commented

kiru() is now the alternative to chop(), instead of kut().

  • 🎉1