clmm + R 4.0 throws warning with long formula

[stragu]

Using R 4.0.2 and ordinal 2019.12-10, I get a warning when using a formula with long names (but not changing the underlying data or the formula structure).

See for example this reprex:

library(ordinal)
wine1 <- wine
# short version:
mm <- clmm(rating ~ temp + contact + (1|judge),
           data = wine1) # no warning
names(wine1)[2] <- "a_very_very_very_very_very_long_name"
# long version:
mm <- clmm(a_very_very_very_very_very_long_name ~ temp + contact + (1|judge),
           data = wine1) # warning
#> Warning: Using formula(x) is deprecated when x is a character vector of length > 1.
#>   Consider formula(paste(x, collapse = " ")) instead.
# it's the length of the whole formula that matters, not the length
# of a single variable's name:
wine2 <- wine
names(wine2) <- lapply(names(wine), paste0, "_quite_long")
mm <- clmm(rating_quite_long ~ temp_quite_long + contact_quite_long + (1|judge_quite_long),
           data = wine2) # warning too
#> Warning: Using formula(x) is deprecated when x is a character vector of length > 1.
#>   Consider formula(paste(x, collapse = " ")) instead.

^{Created on 2020-07-23 by the reprex package (v0.3.0)}

I believe this might be because of how the clmm.formulae() function is defined, and might very well need a fix similar to issue #13 and PR #30, but I have a limited understanding of the functions involved, so I could be completely wrong! 🙂

[ghost]

Just echoing stragu with regard to the warning in ordinal::clmm (and the same setup: R 4.0.2 and ordinal_2019.12-10)

This model

m <- clmm(JUDGMENT ~ 1 + NATIVE_LG * AGREEMENT + EMPHPRON + NEGATION + (1|SUBJECT), data=x)

runs fine but throws a warning message

Using formula(x) is deprecated when x is a character vector of length > 1.
  Consider formula(paste(x, collapse = " ")) instead.`

but if the last predictor is deleted, then it works without problems:

m <- clmm(JUDGMENT ~ 1 + NATIVE_LG * AGREEMENT + EMPHPRON + (1|SUBJECT), data=x)

[runehaubo]

Thanks. Will fix soon. Cheers Rune lør. 15. aug. 2020 kl. 18.46 skrev Stefan Th. Gries < notifications@github.com>:

…

Just echoing stragu with regard to the warning in ordinal::clmm (and the same setup: R 4.0.2 and ordinal_2019.12-10) This model m <- clmm(JUDGMENT ~ 1 + NATIVE_LG * AGREEMENT + EMPHPRON + NEGATION + (1|SUBJECT), data=x) runs fine but throws a warning message Using formula(x) is deprecated when x is a character vector of length > 1. Consider formula(paste(x, collapse = " ")) instead.` but if the last predictor is deleted, then it works without problems: m <- clmm(JUDGMENT ~ 1 + NATIVE_LG * AGREEMENT + EMPHPRON + (1|SUBJECT), data=x) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#34 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABHZZLOIKVJWUMVYE3XD4BLSA23XLANCNFSM4PFJQH5Q> .

[runehaubo]

Now fixed in the version on GitHub dated 2020-08-22.

Thanks again for mentioning this!

[florenceruby]

Hi Rune,
I am experiencing the same issue as above using ordinal_2019.12-10 on R 4.0.2.
Because I have to analyse data on a secure platform we can only access packages from CRAN. Is there a chance the updates you made on the package will be avialble on CRAN soon?

Thanks a lot

[stragu]

Hi @runehaubo
Just wanted to report that a colleague ran into the same warning with clm(), so I assume that function needs a similar fix as well?
Cheers