Comparing grep() vs. grepl() in R: What’s the Difference?

Two functions that people often get mixed up in R are grep() and grepl(). Both functions allow you to see whether a certain pattern exists in a character string, but they return different results:

• grepl() returns TRUE when a pattern exists in a character string.
• grep() returns a vector of indices of the character strings that contain the pattern.

The following example illustrates this difference:

#create a vector of data
data grep('Guard', data)
[1] 1 2

grepl('Guard', data) 
[1]  TRUE  TRUE FALSE FALSE FALSE

The following examples show when you might want to use one of these functions over the other.

When to Use grepl()

1. Filter Rows that Contain a Certain String

One of the most common uses of grepl() is for filtering rows in a data frame that contain a certain string:

library(dplyr)

#create data frame
df #filter rows that contain the string 'Guard' in the player column
df %>% filter(grepl('Guard', player))

   player points rebounds
1 P Guard     12        5
2 S Guard     15        7

Related: How to Filter Rows that Contain a Certain String Using dplyr

When to Use grep()

1. Select Columns that Contain a Certain String

You can use grep() to select columns in a data frame that contain a certain string:

library(dplyr)

#create data frame
df #select columns that contain the string 'p' in their name
df %>% select(grep('p', colnames(df)))

     player points
1   P Guard     12
2   S Guard     15
3 S Forward     19
4 P Forward     22
5    Center     32

2. Count the Number of Rows that Contain a Certain String

You can use grep() to count the number of rows in a data frame that contain a certain string:

#create data frame
df #count how many rows contain the string 'Guard' in the player column
length(grep('Guard', df$player))

[1] 2

You can find more R tutorials here.