Is there an elegant tidyverse method for renaming columns en masse?-CodePudding

I'm working with a small dataset built from a Google Form. The column names are the full survey questions eg.

"1. What team most describes your work?"
"2. Are you a manager?"
"3. How long have you been with x?"

I want to rename all of these columns "1", "2", "3" etc. I know that I can do the below, but I'm hoping there's a more elegant/quicker way. There are 23 columns like this.

survey %>% rename_with(
  `1` = `1. What team most describes your work?`,
  `2` = `2. Are you a manager?`...

This works, but is slow to compose. Anyone have anything better in tidyverse? I'm new to R and coding generally, so any tips appreciated.

CodePudding user response：

You can supply multiple columns in rename_with(.cols = your_columns), and use a function to apply to the selected columns. Here, the question number (the digit before the dot) is captured ((\\d )) and the whole column name is replaced by that digit (regex capture group 1 \\1).

If you omit the .cols argument, all columns will be selected by default.

library(dplyr)

# dummy df
df <- tibble("1. What team most describes your work?" = "1",
             "2. Are you a manager?" = "2",
             "3. How long have you been with x?" = "3")

df %>% rename_with(~sub("(\\d )\\..*$", "\\1", .x))

# A tibble: 1 × 3
  `1`   `2`   `3`  
  <chr> <chr> <chr>
1 1     2     3

CodePudding user response：

There are more concise solutions possible. One of them is with str_extract:

library(stringr)
df %>% rename_with(~str_extract(., "\\d "))

Here we extract the first digit in the names strings.

Another is with str_remove:

df %>% rename_with(~str_remove(., "\\..*"))

Here we remove anything from the period onwards.

Data (thanks to @benson):

df <- tibble("1. What team most describes your work?" = "1",
             "2. Are you a manager?" = "2",
             "3. How long have you been with x?" = "3")