counting a specific variable does not take grouping variables into account

[hlynurhallgrims]

When doing a count on a grouped data.table with dtplyr, only the variable you are counting by is passed to keyby=, but not the current grouping variables (as is the "regular" behaviour in dplyr).

suppressPackageStartupMessages(library(dplyr))

flight_dt <- nycflights13::flights %>% 
  dtplyr::lazy_dt() 

# This is how it works on a data.table
flight_dt %>%   
  group_by(year, month) %>% 
  count(tailnum)
#> Source: local data table [?? x 2]
#> Call:   `_DT1`[, .(n = .N), keyby = .(tailnum)]
#> 
#>   tailnum     n
#>   <chr>   <int>
#> 1 <NA>     2512
#> 2 D942DN      4
#> 3 N0EGMQ    371
#> 4 N10156    153
#> 5 N102UW     48
#> 6 N103US     46
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

# This works and gives the expected output of the previous code
flight_dt %>% 
  group_by(year, month, tailnum) %>% 
  count()
#> Source: local data table [?? x 4]
#> Call:   `_DT1`[, .(n = .N), keyby = .(year, month, tailnum)]
#> 
#>    year month tailnum     n
#>   <int> <int> <chr>   <int>
#> 1  2013     1 <NA>      155
#> 2  2013     1 N0EGMQ     41
#> 3  2013     1 N10156     28
#> 4  2013     1 N102UW      1
#> 5  2013     1 N103US      4
#> 6  2013     1 N104UW      2
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

Created on 2019-08-22 by the reprex package (v0.2.0).

---

p.s. this package is amazing 😍

[hadley]

Slightly simpler reprex:

library(dtplyr)
suppressPackageStartupMessages(library(dplyr))

mtcars %>%  
  group_by(vs) %>% 
  count(am)
#> # A tibble: 4 x 3
#> # Groups:   vs [2]
#>      vs    am     n
#>   <dbl> <dbl> <int>
#> 1     0     0    12
#> 2     0     1     6
#> 3     1     0     7
#> 4     1     1     7

mtcars %>%  
  lazy_dt() %>% 
  group_by(vs) %>% 
  count(am)
#> Source: local data table [?? x 2]
#> Call:   `_DT1`[, .(n = .N), keyby = .(am)]
#> 
#>      am     n
#>   <dbl> <int>
#> 1     0    19
#> 2     1    13
#> 
#> # Use as.data.table()/as.data.frame()/as_tibble() to access results

^{Created on 2019-08-22 by the reprex package (v0.3.0)}