summarize_all() throws error when using tilde notation

[krlmlr]

I'm not sure if this will be fixed with tidyverse/dplyr#3433. Also, I don't understand why the second example gives a warning. Thanks @romanik!

library(tidyverse)
dbplyr::memdb_frame(a = 1, b = 2) %>% 
  summarize_all(~mean(.))
#> Error in mean(.): object 'a' not found
dbplyr::memdb_frame(a = 1, b = 2) %>% 
  summarize_all(mean)
#> Warning: Missing values are always removed in SQL.
#> Use `AVG(x, na.rm = TRUE)` to silence this warning

#> Warning: Missing values are always removed in SQL.
#> Use `AVG(x, na.rm = TRUE)` to silence this warning
#> # Source:   lazy query [?? x 2]
#> # Database: sqlite 3.22.0 [:memory:]
#>       a     b
#>   <dbl> <dbl>
#> 1     1     2
dbplyr::memdb_frame(a = 1, b = 2) %>% 
  summarize_if(funs(is.numeric), ~mean(.))
#> Applying predicate on the first 100 rows
#> Error in mean(.): object 'a' not found

Created on 2018-07-25 by the reprex package (v0.2.0).

[karldw]

Lists of functions seem to have the same problem, and I get the same error for all of the summarize_*, mutate_*, and transmute_* functions.

library(tidyverse)

fn_list <- list(
  p=function(x) x/2,
  q=function(x) x/4
)
df <- dbplyr::memdb_frame(a=1, b=2) 

df %>%
  collect() %>% 
  transmute_at(vars(a,b), .funs=fn_list)
#> A tibble: 1 x 4
#>    a_p   b_p   a_q   b_q
#>  <dbl> <dbl> <dbl> <dbl>
#> 1  0.5     1  0.25   0.5

df %>% transmute_at(vars(a,b), .funs=fn_list)
#> Error in (function (x)  : object 'a' not found

[hadley]

Mild reprex tweaking

library(dplyr, warn.conflicts = FALSE)
mf <- dbplyr::memdb_frame(a = 1, b = 2)

# WORKS
mf %>% summarize_all(mean, na.rm = TRUE)
#> # Source:   lazy query [?? x 2]
#> # Database: sqlite 3.22.0 [:memory:]
#>       a     b
#>   <dbl> <dbl>
#> 1     1     2

# FAILS
mf %>% summarize_all(~mean(.))
#> Error in mean(.): object 'a' not found
mf %>% summarize_all(list(mean))
#> Error in (function (x, ...) : object 'a' not found

^{Created on 2019-01-02 by the reprex package (v0.2.1)}

[hadley]

summarise_all() creates a list of inlined functions so an even more focussed reprex is this:

library(dbplyr)

partial_eval(rlang::expr(mean(x)), vars = "x")
#> mean(x)
partial_eval(rlang::expr((!!mean)(x)), vars = "x")
#> Error in (function (x, ...) : object 'x' not found

^{Created on 2019-02-06 by the reprex package (v0.2.1.9000)}