# "Find and Replace" in Multiple Variable Names

Renaming a variable/set of variables or column names is fairly straightforward. And there are plenty of resources on The Google. If you’re familiar with the dplyr package in R, you’ve probably used select() and rename() a lot. What may not be as straight forward to a beginner or intermediate R user is how to rename a group of variables at the same time or “find and replace” a string of text in a group of variable names—as opposed to making the changes one by one.

This issue has come up most frequently for me in 3 ways:

1. I download a dataset and a group of variables has a simlar “prefix” or “suffix” added to the variable name. E.g. prefix_varname1, prefix_varname2 or varname1_suffix, varname2_suffix.

2. I “collapse” or “summarise” whole or part of a dataset based on some common transformation, like a mean. And I get a new variable/column name like varname1_mean, varname2_mean as an end product.

3. I join/merge two or more large datasets and there are identical variable names in those datasets, and I see something like var1.x or var1.y

Using just select() or rename() you can change these variable names one by one to another name. You might do something like this…

library(tidyverse)
library(stringr)

## CREATE THE FAKE DATA
df <- tribble(
~prefix_x, ~prefix_y, ~prefix_z, # column names
#-----------|---------|---------|
"country1",   100,    1,         # values in each column
"country2",   500,    2
)

df
## # A tibble: 2 x 3
##   prefix_x prefix_y prefix_z
##      <chr>    <dbl>    <dbl>
## 1 country1      100        1
## 2 country2      500        2
## RENAME YOUR VARIABLES
df %>%
rename(x = prefix_x,
y = prefix_y,
z = prefix_z
)
## # A tibble: 2 x 3
##          x     y     z
##      <chr> <dbl> <dbl>
## 1 country1   100     1
## 2 country2   500     2

This is a perfectly good solution. Nothing wrong with using simple tools! This gets time consuming though if you have a lot of variables.

If you’re a beginner, or even intermediate, you may not have read the dplyr documentation thoroughly because there’s so much to consume already 😧. But, if you glanced at it you might have noticed a reference to the “scoped” variants of select() or rename()— namely, select_all, select_if, select_at, rename_all, rename_if, rename_at. In R, type ?select and you’ll see what I’m talking about. These functions are designed to tackle our problem, but I remember there being a bit of a hurdle to understand how to actually use them.

There are many ways to solve this problem, and I’ve included a few examples of how to do this because one way might be easier for you to remember than another. I repeat…there are many!

### Solution 1

Initially, I found this example to be easiest to remember because it didn’t require knowledge of something like “quosure” or purrr. First, we’ll create the data frame. Second, we’ll “find and replace”/delete a string of text in the variable name using a stringr package function str_replace_all. This package and it’s many functions are the tidyverse way to use regex 1 for string/text manipulation.

## FAKE DATA
df2 <- tribble(
~x_mean, ~y_mean, ~z_mean,  # column names
#------|---------|-------|
2.5,    100,     1,        # values in each column
5,      500,     0.5
)

df2
## # A tibble: 2 x 3
##   x_mean y_mean z_mean
##    <dbl>  <dbl>  <dbl>
## 1    2.5    100    1.0
## 2    5.0    500    0.5
## RENAME THE VARS BY TAKING OUT "_mean"
df2 %>%
dplyr::rename_all(
funs(stringr::str_replace_all(., "_mean", ""))
)
## # A tibble: 2 x 3
##       x     y     z
##   <dbl> <dbl> <dbl>
## 1   2.5   100   1.0
## 2   5.0   500   0.5

I’ll explain really quickly what a few parts of the code do because it might help you remember how to write it in the future. . in str_replace_all() is a place holder of sorts for the data frame object, df2, we already created. The next argument in the function is the pattern we want to find in the variable name, i.e. "_mean". The last argument, is what we want to replace it with. In this case we just want to “delete” the variable names’ suffix.

funs() is one way to call an outside tool/function in the “scoped” dplyr functions, and add some expression inside of that called tool/function.

### Solution 2

## RENAME THE VARS BY TAKING OUT "_mean"
df2 %>%
dplyr::rename_all(
~stringr::str_replace_all(., "_mean", "")
)
## # A tibble: 2 x 3
##       x     y     z
##   <dbl> <dbl> <dbl>
## 1   2.5   100   1.0
## 2   5.0   500   0.5

This example uses purrr-style notation. You can read more about it in the purrr documentation.2

### Solution 3

library(purrr)
# or... library(tidyverse) if you loaded that already

## RENAME THE VARS BY TAKING OUT "_mean"
df2 %>%
set_names(~stringr::str_replace_all(., "_mean", "")
)
## # A tibble: 2 x 3
##       x     y     z
##   <dbl> <dbl> <dbl>
## 1   2.5   100   1.0
## 2   5.0   500   0.5

This uses the set_names function in purrr.

Like I mentioned above, there are a lot of different tools out there to do this. Instead of using the stringr package and functions, you might like using base R’s grep functions. Those would work here too by just substituting them for stringr::str_replace_all.

1. “regular expressions”

2. useful sites for reading more if unfamiliar: RStudio’s purrr intro and “jennybc”’s purrr tutorial