0

I see when I run library(multcomp) (see this example), it complains (among other things):

Attaching package: ‘MASS’

The following object is masked from ‘package:dplyr’:

    select

Masking select is going to be awkward.

I suppose I can use multcomp without putting it into the namespace with library.

EDIT: What is the easiest way to run Dunnett's test that works well with the tidyverse?

(Old question: Is there another package that offers Dunnett's test that is more compatible with the tidyverse?)

dfrankow
  • 20,191
  • 41
  • 152
  • 214
  • 2
    If you have a preference for masking, use `library(confliced);conflict_prefer("select", winner = "dplyr", quiet = FALSE)` – akrun May 29 '20 at 23:25
  • 2
    Have you tried loading dplyr/tidyverse *after* multcomp? Then `dplyr::select()` will mask `MASS::select()`, which sounds like it would be preferable to you. That's the advice given in https://stackoverflow.com/a/24202391/8386140 . You could also do `select <- dplyr::select` a la https://stackoverflow.com/a/43051797/8386140 (Apparently conflicts with `dplyr` are a pretty common problem) – duckmayr May 29 '20 at 23:26
  • 1
    Can you please rephrase this to make it about the problem you're trying to solve (`multcomp` loads `MASS` which masks `dplyr::select`) rather than "is there a package that ..."? (which is off-topic for StackOverflow) ... ? – Ben Bolker May 30 '20 at 00:55
  • @BenBolker Thanks for your feedback. Tried rephrasing. The problem I'm trying to solve is not "multcomp loads MASS..". That is a symptom. I am trying to run Dunnett's test smoothly with tidyverse. – dfrankow Jun 02 '20 at 15:22

0 Answers0