Unlike other statistical packages, R has a robust and simple to use set of string manipulation functions. These functions become useful in a number of situations, including: dynamically creating variables, generating tabular and graphical output, reading and writing from text files and the web, and managing character data (e.g., recoding free response or other character data). This tutorial walks through some of the basic string manipulations functions.
The simplest and most important string manipulation function is paste. It allows the user to concatenate character strings (and vectors of character strings) in a number of different ways.
The easiest way to use paste is simply to concatenate several values together:
paste("1", "2", "3")
## [1] "1 2 3"
The result is a single string (i.e., one-element character vector) with the numbers separated by spaces (which is the default). We can also separate by other values:
paste("1", "2", "3", sep = ",")
## [1] "1,2,3"
A helpful feature of paste is that it coerces objects to character before concatenating, so we can get the same result above using:
paste(1, 2, 3, sep = ",")
## [1] "1,2,3"
This also means we can combine objects of different classes (e.g., character and numeric):
paste("a", 1, "b", 2, sep = ":")
## [1] "a:1:b:2"
Another helpful feature of paste is that it is vectorized, meaning that we can concatenate each element of two or more vectors in a single call:
a <- letters[1:10]
b <- 1:10
paste(a, b, sep = "")
## [1] "a1" "b2" "c3" "d4" "e5" "f6" "g7" "h8" "i9" "j10"
The result is a 10-element vector, where the first element of a has been pasted to the first element of b and so forth.
We might want to collapse a multi-item vector into a single string and for this we can use the collapse argument to paste:
paste(a, collapse = "")
## [1] "abcdefghij"
Here, all of the elements of a are concatenated into a single string.
We can also combine the sep and collapse arguments to obtain different results:
paste(a, b, sep = "", collapse = ",")
## [1] "a1,b2,c3,d4,e5,f6,g7,h8,i9,j10"
paste(a, b, sep = ",", collapse = ";")
## [1] "a,1;b,2;c,3;d,4;e,5;f,6;g,7;h,8;i,9;j,10"
The first result above concatenates corresponding elements from each vector without a space and then separates them by a comma. The second result concatenates corresponding elements with a comma between the elements and separates each pair of elements by semicolon.
The strsplit function offers essentially the reversal of paste, by cutting a string into parts based on a separator. Here we can collapse our a vector and then split it back into a vector:
a1 <- paste(a, collapse = ",")
a1
## [1] "a,b,c,d,e,f,g,h,i,j"
strsplit(a1, ",")
## [[1]]
## [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j"
Note: strsplit returns a list of results, so accessing the elements requires using the [[]] (double bracket) operator. To get the second element from the split vector, use:
strsplit(a1, ",")[[1]][2]
## [1] "b"
The reason for this return value is that strsplit is also vectorized. So we can split multiple elements of a character vectors in one call:
b1 <- paste(a, b, sep = ",")
b1
## [1] "a,1" "b,2" "c,3" "d,4" "e,5" "f,6" "g,7" "h,8" "i,9" "j,10"
strsplit(b1, ",")
## [[1]]
## [1] "a" "1"
##
## [[2]]
## [1] "b" "2"
##
## [[3]]
## [1] "c" "3"
##
## [[4]]
## [1] "d" "4"
##
## [[5]]
## [1] "e" "5"
##
## [[6]]
## [1] "f" "6"
##
## [[7]]
## [1] "g" "7"
##
## [[8]]
## [1] "h" "8"
##
## [[9]]
## [1] "i" "9"
##
## [[10]]
## [1] "j" "10"
The result is a list of split vectors.
Sometimes we want to get every single character from a character string, and for this we can use an empty separator:
strsplit(a1, "")[[1]]
## [1] "a" "," "b" "," "c" "," "d" "," "e" "," "f" "," "g" "," "h" "," "i"
## [18] "," "j"
The result is every letter and every separator split apart.
strsplit also supports much more advanced character splitting using “regular expressions.” We address that in a separate tutorial.
Sometimes we want to know how many characters are in a string, or just get a subset of the characters. R provides two functions to help us with these operations: nchar and substring.
You can think of nchar as analogous to length but instead of telling you how many elements are in a vector it tells you how many characters are in a string:
d <- "abcd"
length(d)
## [1] 1
nchar(d)
## [1] 4
nchar is vectorized, which means we can retrieve the number of characters in each element of a characte vector in one call:
e <- c("abc", "defg", "hi", "jklmnop")
nchar(e)
## [1] 3 4 2 7
substring lets you extract a part of a string based on the position of characters in the string and can be combined with nchar:
f <- "hello"
substring(f, 1, 1)
## [1] "h"
substring(f, 2, nchar(f))
## [1] "ello"
substring is also vectorized. For example we could extract the first character from each element of a vector:
substring(e, 1, 1)
## [1] "a" "d" "h" "j"
Or even the last character of elements with different numbers of characters:
e
## [1] "abc" "defg" "hi" "jklmnop"
nchar(e)
## [1] 3 4 2 7
substring(e, nchar(e), nchar(e))
## [1] "c" "g" "i" "p"