class is not respected in a grouped mutate

[bnicenboim]

Hi, I actually know why this is happening, and not sure if it's fixable (without slowing down the grouped mutate). But if not, it's worth a message or a warning when it happens.

suppressMessages(library(tidytable))
df <- data.frame(a= 1:10, b= 1:10, c= 1:10)
# class is not respected here
df %>% 
  mutate.(a = a > 5, .by = "c")
#> # A tidytable: 10 × 3
#>        a     b     c
#>    <int> <int> <int>
#>  1     0     1     1
#>  2     0     2     2
#>  3     0     3     3
#>  4     0     4     4
#>  5     0     5     5
#>  6     1     6     6
#>  7     1     7     7
#>  8     1     8     8
#>  9     1     9     9
#> 10     1    10    10
# class is respected here:
df %>% 
  mutate.(a = a > 5)
#> # A tidytable: 10 × 3
#>    a         b     c
#>    <lgl> <int> <int>
#>  1 FALSE     1     1
#>  2 FALSE     2     2
#>  3 FALSE     3     3
#>  4 FALSE     4     4
#>  5 FALSE     5     5
#>  6 TRUE      6     6
#>  7 TRUE      7     7
#>  8 TRUE      8     8
#>  9 TRUE      9     9
#> 10 TRUE     10    10

^{Created on 2022-08-31 with reprex v2.0.2}

I do get a warning if I do this (which also doesn't work)

df %>% 
  mutate.(a = as.character(a), .by = "c")

[markfairbanks]

Unfortunately this isn't fixable by me, but if you want you could open an issue in the data.table repo. The actual computation is done by data.table, so however they handle it is how it'll work.

[bnicenboim]

mmh, yeah, but given that data.table works by reference, it's clearer what's going on...

You can still do what I intended with data.table like this:

dt[,.(a = a > 5), by = c][]

(In any case, as I said I kind of assumed this wasn't fixable...)