all.equal on object with NULL names causes 'Error: not compatible with STRSXP' -- bug or expected?

dplyr, r

Solution

As @Joran points out, this seems to be related to `dplyr`. Filed as an issue: https://github.com/hadley/dplyr/issues/219

Temporary work around (for my need at least. Will not work for all) is to use

 all.equal.default(x, x)

FYI:

  ## STARTING FROM A FRESH SESSION:

  set.seed(1)
  x <- data.frame(LETTERS[1:3], rnorm(3))
  names(x) <- NULL

  all.equal(x, x)
  # [1] TRUE

  ## Load in dplyr
  library(dplyr)
  all.equal(x, x)
  # Error: not compatible with STRSXP

Problem

In the example below, when names are set to `NULL`, `all.equal` throws `'Error: not compatible with STRSXP'` However, if names are set to `NA` (or some other value), `all.equal` works as normal. Is this expected behavior or a bug? ``` ## SAMPLE DATA set.seed(1) x <- data.frame(LETTERS[1:3], rnorm(3)) names(x) <- NULL x # NA NA # 1 A -0.626454 # 2 B 0.183643 # 3 C -0.835629 all.equal(x, x) # Error: not compatible with STRSXP # add names back in, even 'NA' names(x) <- c(NA, NA) all.equal(x, x) # [1] TRUE ```

Original source