Dplyr "rename_with" giving unique column name error but function creates unique outputs-CodePudding

I created a function that creates unique column names from the existing ones: renameCol.

If I manually create a vector of new column names using that function I can manually set those as the new column names. However, if I use that function in rename_with I get an error about unique names.

library(tidyverse)
renameCol = function(colname)
{
  match = str_match_all(colname, "HealthcareProvider((TaxonomyCode|PrimaryTaxonomySwitch))_([0-9] )")[[1]]
  
  coltype = match[[3]]
  coltype = str_remove(coltype, "(Taxonomy|PrimaryTaxonomy)")
  number = match[[4]]
  
  return(paste0(coltype, "_", number))
}

renameCol("HealthcareProviderPrimaryTaxonomySwitch_11")
#> [1] "Switch_11"
renameCol("HealthcareProviderTaxonomyCode_11")
#> [1] "Code_11"

tb = tibble(
  HealthcareProviderPrimaryTaxonomySwitch_11 = 1,
  HealthcareProviderTaxonomyCode_3 = 2,
  HealthcareProviderPrimaryTaxonomySwitch_9 = 3,
  HealthcareProviderTaxonomyCode_13 = 4
)

tb %>% rename_with(renameCol)
#> Error in `rename_with()`:
#> ! Names must be unique.
#> x These names are duplicated:
#>   * "Switch_11" at locations 1, 2, 3, and 4.

new_colnames = colnames(tb) %>% sapply(renameCol, USE.NAMES = F)
new_colnames
#> [1] "Switch_11" "Code_3"    "Switch_9"  "Code_13"

colnames(tb) = new_colnames
tb
#> # A tibble: 1 x 4
#>   Switch_11 Code_3 Switch_9 Code_13
#>       <dbl>  <dbl>    <dbl>   <dbl>
#> 1         1      2        3       4

^{Created on 2022-06-16 by the reprex package (v2.0.1)}

CodePudding user response：

The answer is present in your question itself. Your function is not vectorised. It works for only one column name at a time.

library(tidyverse)

names(tb)
#[1] "HealthcareProviderPrimaryTaxonomySwitch_11"
#[2] "HealthcareProviderTaxonomyCode_3"          
#[3] "HealthcareProviderPrimaryTaxonomySwitch_9" 
#[4] "HealthcareProviderTaxonomyCode_13"  
     
renameCol(names(tb))
#[1] "Switch_11"

Hence you have to use sapply to make it work for all the columns. rename_with is not a loop (like sapply) so to make it work you can do -

tb %>% rename_with(~sapply(., renameCol))

# A tibble: 1 × 4
#  Switch_11 Code_3 Switch_9 Code_13
#      <dbl>  <dbl>    <dbl>   <dbl>
#1         1      2        3       4

Or change the function to work with more than one column name.

renameCol = function(colname)
{
  match = str_match_all(colname, "HealthcareProvider((TaxonomyCode|PrimaryTaxonomySwitch))_([0-9] )")
  match_data <- do.call(rbind, match)
  coltype = match_data[, 3]
  coltype = str_remove(coltype, "(Taxonomy|PrimaryTaxonomy)")
  number = match_data[, 4]
  
  return(paste0(coltype, "_", number))
}

renameCol(names(tb))

#[1] "Switch_11" "Code_3"    "Switch_9"  "Code_13"  

tb %>% rename_with(renameCol)

# A tibble: 1 × 4
#  Switch_11 Code_3 Switch_9 Code_13
#      <dbl>  <dbl>    <dbl>   <dbl>
#1         1      2        3       4