Extract last word in string in R

Question

What's the most elegant way to extract the last word in a sentence string?

The sentence does not end with a "." Words are seperated by blanks.

sentence <- "The quick brown fox"
TheFunction(sentence)

should return: "fox"

I do not want to use a package if a simple solution is possible. If a simple solution based on package exists, that is also fine.

Give us some details about your data structure, otherwise, we're just all stabbing in the dark while trying to help you out. — A5C1D2H2I1M1N2O1R2T1, Jul 15 '13 at 15:43
What have you tried? I'd imagine you can just reverse the string and separate by `" "`. — Señor O, Jul 15 '13 at 15:43
Related: http://stackoverflow.com/questions/13093931/r-remove-last-word-from-string — Señor O, Jul 15 '13 at 16:01

leo · Answer 1 · 2015-01-20T19:43:10.823

69

Just for completeness: The library stringr contains a function for exactly this problem.

library(stringr)

sentence <- "The quick brown fox"
word(sentence,-1)
[1] "fox"

edited Jan 20 '15 at 19:43

answered Sep 21 '13 at 13:30

leo

score 31 · Accepted Answer · answered Jul 15 '13 at 15:55

31

tail(strsplit('this is a sentence',split=" ")[[1]],1)

Basically as suggested by @Señor O.

answered Jul 15 '13 at 15:55

Roland

It's hard for me to call a method with `[[` elegant :) but that might be my personal feeling for R's lists – JelenaČuklina Dec 28 '16 at 10:14

score 18 · Answer 3 · answered Jul 15 '13 at 15:51

18

x <- 'The quick brown fox'
sub('^.* ([[:alnum:]]+)$', '\\1', x)

That will catch the last string of numbers and characters before then end of the string.

You can also use the regexec and regmatches functions, but I find sub cleaner:

m <- regexec('^.* ([[:alnum:]]+)$', x)
regmatches(x, m)

See ?regex and ?sub for more info.

answered Jul 15 '13 at 15:51

Justin

I'd say Roland's solution to this specific problem is much simpler, but your code provides a good template for doing all sorts of targeted extractions. – Carl Witthoft Jul 15 '13 at 17:20
if you're looking to make the syntax even shorter, you can omit `^` (`sub` will start at the beginning on its own) and replace `[[:alnum:]]` with `\\w` – eddi Jul 15 '13 at 17:23
@eddi Good point. However, I like to be as explicit as possible with regular expressions and only generalize as I find cases I miss rather than trying to find cases where I've over matched. – Justin Jul 15 '13 at 17:37
1

how would the regex be like if we extract the last N words, not just the last one ? – Fagui Curtain Nov 23 '15 at 01:34
Thank you for providing a clean, base R answer. I want to learn regex, not just some clever wrapper for it!! – emilBeBri Mar 26 '18 at 13:02

Rich Scriven · Answer 4 · 2016-08-23T16:12:51.767

13

Another packaged option is stri_extract_last_words() from the stringi package

library(stringi)

stri_extract_last_words("The quick brown fox")
# [1] "fox"

The function also removes any punctuation that may be at the end of the sentence.

stri_extract_last_words("The quick brown fox? ...")
# [1] "fox"

edited Aug 23 '16 at 16:12

answered Jan 20 '15 at 20:34

Rich Scriven

score 5 · Answer 5 · answered Jul 15 '13 at 17:32

5

Going in the package direction, this is the simplest answer I can think of:

library(stringr)

x <- 'The quick brown fox'
str_extract(x, '\\w+$')
#[1] "fox"

answered Jul 15 '13 at 17:32

eddi

5 Answers5