If I have a vector, new_cols and I want to add n number of columns to a data.frame with constant values. But I can't quite grok a tidyverse way of doing this. Here is what I've tried so far:
library(dplyr)
library(purrr)
# Create a tibble
df <- tibble(
col1 = c(1, 2),
col2 = c(3, 4)
)
# Vector of column names
new_cols <- c("miss1", "miss2")
# Using mutate() and !! operator and purrr
map_df(new_cols, ~ {
df %>% mutate(!!.x := "unknown")
})
#> # A tibble: 4 × 4
#> col1 col2 miss1 miss2
#> <dbl> <dbl> <chr> <chr>
#> 1 1 3 unknown <NA>
#> 2 2 4 unknown <NA>
#> 3 1 3 <NA> unknown
#> 4 2 4 <NA> unknown
# Using mutate() and across() function
df <- df %>%
mutate(across(all_of(new_cols), ~"unknown", .names = "{.col}"))
#> Error in `mutate()`:
#> ℹ In argument: `across(all_of(new_cols), ~"unknown", .names =
#> "{.col}")`.
#> Caused by error in `all_of()`:
#> ! Can't subset columns that don't exist.
#> ✖ Columns `miss1` and `miss2` don't exist.
#> Backtrace:
#> ▆
#> 1. ├─df %>% mutate(across(all_of(new_cols), ~"unknown", .names = "{.col}"))
#> 2. ├─dplyr::mutate(., across(all_of(new_cols), ~"unknown", .names = "{.col}"))
#> 3. ├─dplyr:::mutate.data.frame(...)
#> 4. │ └─dplyr:::mutate_cols(.data, dplyr_quosures(...), by)
#> 5. │ ├─base::withCallingHandlers(...)
#> 6. │ └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
#> 7. │ └─dplyr:::expand_across(dot)
#> 8. │ └─dplyr:::across_setup(...)
#> 9. │ └─tidyselect::eval_select(cols, data = data, error_call = error_call)
#> 10. │ └─tidyselect:::eval_select_impl(...)
#> 11. │ ├─tidyselect:::with_subscript_errors(...)
#> 12. │ │ └─rlang::try_fetch(...)
#> 13. │ │ └─base::withCallingHandlers(...)
#> 14. │ └─tidyselect:::vars_select_eval(...)
#> 15. │ └─tidyselect:::walk_data_tree(expr, data_mask, context_mask)
#> 16. │ └─tidyselect:::eval_context(expr, context_mask, call = error_call)
#> 17. │ ├─tidyselect:::with_chained_errors(...)
#> 18. │ │ └─rlang::try_fetch(...)
#> 19. │ │ ├─base::tryCatch(...)
#> 20. │ │ │ └─base (local) tryCatchList(expr, classes, parentenv, handlers)
#> 21. │ │ │ └─base (local) tryCatchOne(expr, names, parentenv, handlers[[1L]])
#> 22. │ │ │ └─base (local) doTryCatch(return(expr), name, parentenv, handler)
#> 23. │ │ └─base::withCallingHandlers(...)
#> 24. │ └─rlang::eval_tidy(as_quosure(expr, env), context_mask)
#> 25. ├─tidyselect::all_of(new_cols)
#> 26. │ └─tidyselect:::as_indices_impl(x, vars = vars, strict = TRUE)
#> 27. │ └─tidyselect:::chr_as_locations(x, vars, call = call, arg = arg)
#> 28. │ └─vctrs::vec_as_location(...)
#> 29. └─vctrs (local) `<fn>`()
#> 30. └─vctrs:::stop_subscript_oob(...)
#> 31. └─vctrs:::stop_subscript(...)
#> 32. └─rlang::abort(...)
This is actually what I am after but I want it generalizable because it could take any number of columns.
# Using mutate() and !! operator
df %>%
mutate(
!!new_cols[1] := "unknown", # Add first column with NA values
!!new_cols[2] := "unknown"
) # Add second column with NA values
#> # A tibble: 2 × 4
#> col1 col2 miss1 miss2
#> <dbl> <dbl> <chr> <chr>
#> 1 1 3 unknown unknown
#> 2 2 4 unknown unknown
Created on 2023-07-07 with reprex v2.0.2
