Explicit namespacing for translated functions

[TimTaylor]

In version v0.3.2 functions such as wday() seemed to work without needing lubridate loaded or explicit namespacing (i.e. lubridate::wday()). This no longer seems to be the case. Is this deliberate?

I've noticed this difference with both the main branch (0.99.99.9923) and the 0.4.1 release.

[krlmlr]

Thanks. This is by design, it works with library(lubridate) .

We need to make the error a bit nicer, though.

options(conflicts.policy = list(warn = FALSE))
library(dplyr)

# Bad:
duckplyr::duckdb_tibble(date = Sys.Date() + 1:5, .funnel = TRUE) |>
  mutate(dow = wday(date))
#> Error in `mutate()`:
#> ! This operation cannot be carried out by DuckDB, and the input is a
#>   funneled duckplyr frame.
#> ℹ Use `compute(funnel = FALSE)` to materialize to temporary storage and
#>   continue with duckplyr.
#> ℹ See `vignette("funnel")` for other options.
#> Caused by error in `rel_find_call()` at duckplyr/R/translate.R:162:3:
#> ! Function `wday` does not map to `lubridate::wday`.

# Good:
duckplyr::duckdb_tibble(date = Sys.Date() + 1:5, .funnel = TRUE) |>
  mutate(dow = lubridate::wday(date))
#> # A duckplyr data frame: 2 variables
#>   date         dow
#>   <date>     <int>
#> 1 2025-01-30     5
#> 2 2025-01-31     6
#> 3 2025-02-01     7
#> 4 2025-02-02     1
#> 5 2025-02-03     2

# Also good:
library(lubridate)
duckplyr::duckdb_tibble(date = Sys.Date() + 1:5, .funnel = TRUE) |>
  mutate(dow = wday(date))
#> # A duckplyr data frame: 2 variables
#>   date         dow
#>   <date>     <int>
#> 1 2025-01-30     5
#> 2 2025-01-31     6
#> 3 2025-02-01     7
#> 4 2025-02-02     1
#> 5 2025-02-03     2

^{Created on 2025-01-29 with reprex v2.1.1}