lead / lag within function with .data$

[shriv]

I'm using dtplyr within an R package of data transformations. I'm using dtplyr_1.3.1 and dplyr_1.1.1 on an M1 mac.

I rely heavily on lead() and lag(). However, I find that neither lead() nor lag() can be used with columns referenced with .data$<column_name>. test_fn_working() has no issues but if you run test_fun_not_working() from the reprex, you get the following error.

Is there a workaround to use .data$<column_name> within lead() and lag() functions?

Error:
! Can't subset `.data` outside of a data mask context.
Run `rlang::last_trace()` to see where the error occurred.

library(dplyr)
library(dtplyr)

test_fn_working <- function(){
  
  dplyr::tibble(
    a = seq(1, 10, 1)
  ) |> 
    dtplyr::lazy_dt() |> 
    dplyr::mutate(
      b = lead(a)
    )
}

test_fn_not_working <- function(){
  
  dplyr::tibble(
    a = seq(1, 10, 1)
  ) |> 
    dtplyr::lazy_dt() |> 
    dplyr::mutate(
      b = lead(.data$a)
    )
}

[shriv]

Hi @markfairbanks and @eutwt - thanks so much for merging the PR for this bug so quickly! :-) Much appreciated..! Do you know when you might release a dev version with this fix? Thanks!

[markfairbanks]

You're welcome, thanks for letting us know about the issue 😄

You can install the dev version that contains the fix with this command:

devtools::install_github("tidyverse/dtplyr")

As for the next CRAN release, I would guess it will be in a few months.

[shriv]

Fab! I'll install the github dev version to get the fix :-)