Reusing column lists in filter and sort?

[not-my-profile]

Can the following query be rewritten without having to repeat the columns?

from foo
filter a != null
filter b != null
filter c != null
sort {a, b, c}

[eitsupi]

Maybe related to #2386

[aljazerzen]

Does the tidyverse have a way to do this?

[eitsupi]

It is not possible to do that directly, but it avoids repetitive statements if there are rules for column names, if all columns are to be selected, or if combinations of columns can be predefined.

df <- tibble::tribble(
  ~x, ~y, ~z,
  NA, 1, "c",
  2, NA, "b",
  5, 3, "a",
  4, 4, "d",
  3, 5, NA
)

cols <- c("x", "y")

df |>
  dplyr::filter(dplyr::if_all(all_of(cols), \(col) !is.na(col))) |>
  dplyr::arrange(dplyr::across(all_of(cols)))
#> # A tibble: 3 × 3
#>       x     y z
#>   <dbl> <dbl> <chr>
#> 1     3     5 <NA>
#> 2     4     4 d
#> 3     5     3 a

df |>
  dplyr::filter(dplyr::if_all(everything(), \(col) !is.na(col))) |>
  dplyr::arrange(dplyr::across(everything()))
#> # A tibble: 2 × 3
#>       x     y z
#>   <dbl> <dbl> <chr>
#> 1     4     4 d
#> 2     5     3 a

^{Created on 2023-07-26 with reprex v2.0.2}

[aljazerzen]

I don't think that's possible, even with some complicated functions.