rename_with issue

[linuxiaobai]

---

Looks like from the version v1.1.0, the function rename_with will get error if no variables are selected. In my case, I have data from many spreedsheets, and I need to rename variable ABC to AAA if variable ABC is available. If ABC doesn't exist in one spreadsheet, then nothing to do. Actually I have several variables like ABC. If I have to judge if all those variables exist in all the spreedsheets, the logic will be quality complex. So I have tired below. It seems rename_with has a bug for the last piece of code.

# library
library(dplyr)

test=iris
names(test)

# this code works for the verion lower than v1.1.0
test0<-test %>% rename_with("newname",starts_with("Spec2ie"))

# this code works and no variable name changes
test1<-test %>% rename_with(~character(0),starts_with("Spec2ie"))
names(test1)

# this code don't work
test2<-test %>% rename_with(~ifelse(length(.)==0,character(0),"newname"),
                        starts_with("Spec2ie"))
# Error in `rename_with()`:
# ! `.fn` must return a vector of length 0, not 1.
# Run `rlang::last_trace()` to see where the error occurred.

[manhnguyen48]

Couple of ways to achieve what you want but generally you want a lookup vector to control how the renaming happens:

library(dplyr, warn.conflicts = FALSE)

name_lookup <- c("AAA" = "ABC",
                 "AA1" = "AAC")
test <- tibble::tibble(ABC = 1:10,
                       AAC = 11:20,
                       BBB = 21:30)
#If we don't need to specify the specific groups of columns where renaming happens,
#we can use dplyr::rename
dplyr::rename(test, all_of(name_lookup))
#> # A tibble: 10 × 3
#>      AAA   AA1   BBB
#>    <int> <int> <int>
#>  1     1    11    21
#>  2     2    12    22
#>  3     3    13    23
#>  4     4    14    24
#>  5     5    15    25
#>  6     6    16    26
#>  7     7    17    27
#>  8     8    18    28
#>  9     9    19    29
#> 10    10    20    30

#dplyr::rename_with requires a function
#stringr::str_replace_all takes in a named vector but the order is reverse
#i.e. c(pattern = replacement)
dplyr::rename_with(test,
                   .fn = ~ stringr::str_replace_all(.x,
                                                    setNames(names(name_lookup),
                                                             unlist(name_lookup))), 
                   .cols = starts_with("A"))
#> # A tibble: 10 × 3
#>      AAA   AA1   BBB
#>    <int> <int> <int>
#>  1     1    11    21
#>  2     2    12    22
#>  3     3    13    23
#>  4     4    14    24
#>  5     5    15    25
#>  6     6    16    26
#>  7     7    17    27
#>  8     8    18    28
#>  9     9    19    29
#> 10    10    20    30

#If the specified .cols is not covered in .fn then we won't see any changes
dplyr::rename_with(test,
                   .fn = ~ stringr::str_replace_all(.x,
                                                    setNames(names(name_lookup),
                                                             unlist(name_lookup))), 
                   .cols = starts_with("B"))
#> # A tibble: 10 × 3
#>      ABC   AAC   BBB
#>    <int> <int> <int>
#>  1     1    11    21
#>  2     2    12    22
#>  3     3    13    23
#>  4     4    14    24
#>  5     5    15    25
#>  6     6    16    26
#>  7     7    17    27
#>  8     8    18    28
#>  9     9    19    29
#> 10    10    20    30

^{Created on 2023-06-26 with reprex v2.0.2}

[hadley]

I don't think you need rename_with() for this use case, you can just use rename():

iris %>% rename(newname = starts_with("Spec2ie"))

Additionally, ifelse() is vectorised so if you did want to do this with rename_with(), you need to use if:

iris %>% rename_with(\(nms) if (length(nms)==0) character(0) else "newname", starts_with("Spec2ie"))

[linuxiaobai]

I don't think you need rename_with() for this use case, you can just use rename():

iris %>% rename(newname = starts_with("Spec2ie"))
Additionally, ifelse() is vectorised so if you did want to do this with rename_with(), you need to use if:

iris %>% rename_with((nms) if (length(nms)==0) character(0) else "newname", starts_with("Spec2ie"))

Thank you very much for your reply! I did need to use rename_with as I need to use the function instead of renaming to a single value. I refer to your solution and the problems was resolved!