We have been exploring various R packages for handling text for natural language processing. In this twenty-third article in the R, Statistics and Machine Learning series, we delve into the ‘stringr’ package, which provides a comprehensive set of functions to easily work with strings.
We will use R version 4.1.2 installed on Parabola GNU/Linux-libre (x86-64) for the code snippets.
$ R --versionR version 4.1.2 (2021-11-01) -- “Bird Hippie”Copyright (C) 2021 The R Foundation for Statistical ComputingPlatform: x86_64-pc-linux-gnu (64-bit)R is free software and comes with ABSOLUTELY NO WARRANTY.You are welcome to redistribute it under certain conditions.Type ‘license()’ or ‘licence()’ for distribution details. |
You can install and load the ‘stringr’ package using the following commands:
> install.packages(“stringr”)Installing package into ‘/usr/local/lib/R/site-library’(as ‘lib’ is unspecified)...** installing vignettes** testing if installed package can be loaded from temporary location** testing if installed package can be loaded from final location** testing if installed package keeps a record of temporary installation path* DONE (stringr)> library(stringr) |
str_to_
The str_to_() functions provide methods to transform strings to upper case and lower case, format titles, and convert text into a sentence format. The syntax for this function is as follows:
str_to_<function>(string, locale = “en”) |
A few examples are given below:
> t <- “R, Statistics and Machine Learning”> str_to_upper(t)[1] “R, STATISTICS AND MACHINE LEARNING”> str_to_lower(t)[1] “r, statistics and machine learning”> str_to_title(t)[1] “R, Statistics And Machine Learning”> str_to_sentence(t)[1] “R, statistics and machine learning” |
str_count
You can count the number of occurrences of a character in a string with the str_count() function. The syntax usage is as follows:
str_count(string, pattern = “”) |
A couple of examples are given below for reference:
> str_count(t, “a”)[1] 4> str_count(t, c(“a”, “e”))[1] 4 2 |
str_dup
You can duplicate a string with the str_dup() function, which accepts an input string and a number for replication. The number of times for duplication can also be a list as shown below:
> str_dup(t, 1)[1] “R, Statistics and Machine Learning”> str_dup(t, 2)[1] “R, Statistics and Machine LearningR, Statistics and Machine Learning”> str_dup(t, 1:3)[1] “R, Statistics and Machine Learning”[2] “R, Statistics and Machine LearningR, Statistics and Machine Learning”[3] “R, Statistics and Machine LearningR, Statistics and Machine LearningR, Statistics and Machine Learning” |
str_detect
The str_detect() function returns a TRUE Boolean value if the pattern match exists in the given input string, and FALSE otherwise. You can use the regex() syntax for specific patterns for a match. The negate argument, if set to TRUE, can return non-matching elements. Examples to demonstrate this function are given below:
> str_detect(t, “a”)[1] TRUE> str_detect(t, “[ae]”)[1] TRUE> str_detect(t, “^s”)[1] FALSE> str_detect(t, “g$”)[1] TRUEstr_conv |
The str_conv() function can help convert the encoding of a string from the default format. The syntax usage is as follows:
str_conv(string, encoding) |
Examples that use the ISO-8859-1 encoding are given below for reference:
> str_conv(“\xa9”, “ISO-8859-1”)[1] “©”> str_conv(“\xbc”, “ISO-8859-1”)[1] “¼”> str_conv(“\xbd”, “ISO-8859-1”)[1] “½”> str_conv(“\xbe”, “ISO-8859-1”)[1] “¾” |
str_equal
You can compare if two strings are equal using the Unicode rules with the str_equal() function. It accepts the following arguments:
| Argument | Description |
| x | A character vector |
| y | Another character vector |
| locale | ‘en’ for English |
| ignore_case | Boolean value to ignore case |
A couple of examples are shown below:
> str_equal(“hello”, “hi”)[1] FALSE> str_equal(“\u1342”, “\u1342”)[1] TRUE |
str_like
The pattern matching for a string for the SQL LIKE operator syntax is implemented with the str_like() function. The syntax usage is as follows:
str_like(string, pattern, ignore_case) |
A few examples are given below:
> str_like(vowels, “a”)[1] TRUE FALSE FALSE FALSE FALSE> str_like(t, “Mach”)[1] FALSE> str_like(t, “%R%”)[1] TRUEstr_match |
The str_match() function does pattern matching as described in vignette (‘regular-expressions’) and as implemented by string. A couple of examples are given below:
> str_match(t, “[a-z]+”)[,1][1,] “tatistics”> str_match(t, “[a-zA-Z]+”)[,1][1,] “R” |
str_extract
The str_extract() function matches a pattern in a string, and obtains the same. It accepts the following arguments:
| Argument | Description |
| string | Input vector |
| pattern | Regular expression |
| group | Return specified matched group |
| simplify | TRUE returns character matrix FALSE returns list of character vectors |
A few examples are given below:
> str_extract(t, “\\d”)[1] NA> str_extract(t, “and”)[1] “and”> str_extract(t, “[a-z]+”)[1] “tatistics”> str_extract(t, “[a-zA-Z]+”)[1] “R” |
str_flatten
You can convert a character vector to a string using the str_flatten() function. It takes the following arguments:
| Argument | Description |
| string | Input vector |
| collapse | String to insert between elements |
| last | Optional string for the final separator |
| na.rm | Boolean to handle missing values |
A few examples given below illustrate this function:
> vowels <- c(“a”, “e”, “i”, “o”, “u”)> str_flatten(vowels)[1] “aeiou”> str_flatten(vowels[1:3], “-”)[1] “a-e-i”> str_flatten_comma(vowels)[1] “a, e, i, o, u” |
str_locate
The str_locate() function returns the beginning and end position of the first pattern match for a given input string. The str_locate_all() function returns all matching occurrences. A few examples are given below:
> str_locate(t, “a”)start end[1,] 6 6> str_locate(t, “$”)start end[1,] 35 34> str_locate_all(t, “a”)[[1]]start end[1,] 6 6[2,] 15 15[3,] 20 20[4,] 29 29 |
str_sort
You can order, rank or sort a character vector using the str_sort() function. It takes the following arguments:
| Argument | Description |
| x | Character vector |
| decreasing | TRUE for highest to lowest FALSE otherwise (default) |
| na_last | Boolean to handle NA values |
| numeric | Sort digits numerically |
A couple of examples are given below:
> str_sort(vowels)[1] “a” “e” “i” “o” “u”> f <- c(“beta”, “alpha”, “gamma”, “delta”)> str_sort(f)[1] “alpha” “beta” “delta” “gamma” |
str_remove
The str_remove() function removes text that matches a pattern for an input string. It accepts two arguments — an input vector string, and a pattern. An example of the use of this function is given below:
> str_remove(t, “\\s”)[1] “R,Statistics and Machine Learning”> str_remove(t, “[aeiou]”)[1] “R, Sttistics and Machine Learning” |
str_replace
The str_replace() function replaces the first occurrence of the pattern with the replacement string. The syntax usage is as follows:
str_replace(string, pattern, replacement) |
A couple of examples are as follows:
> str_replace(t, “\\s”, “-”)[1] “R,-Statistics and Machine Learning”> str_replace(t, “[aeiou]”, “ “)[1] “R, St tistics and Machine Learning” |
str_split
You can split a string into multiple segments using the str_split() function. It has multiple options.
The str_split() accepts a character vector and returns a list as shown below:
> str_split(t, “ “)[[1]][1] “R,” “Statistics” “and” “Machine” “Learning” |
The str_split_1() uses a single string and returns a character vector. For example:
> str_split_1(t, “and”)[1] “R, Statistics “ “ Machine Learning” |
The str_split_fixed() accepts a character vector and returns a matrix of values. A couple of examples are given below:
> str_split_fixed(t, “ “, 2)[,1] [,2][1,] “R,” “Statistics and Machine Learning”> str_split_fixed(t, “ “, 3)[,1] [,2] [,3][1,] “R,” “Statistics” “and Machine Learning” |
The str_split_i() takes a character vector and returns a character vector. A few examples are given below to demonstrate this function:
> str_split_i(t, “ “, 1)[1] “R,”> str_split_i(t, “ “, 2)[1] “Statistics”> str_split_i(t, “ “, 3)[1] “and” |
You are encouraged to read the stringr reference manual to learn more about its functions, arguments, options, and usage.
