Getting only matched part of the string in R

Is there a function in R that matches regexp and returns only the matched parts? Something like grep -o, so:

> ogrep('.b.',c('abc','开发者_开发知识库1b2b3b4'))
[[1]]
[1] abc

[[2]]
[1] 1b2 3b4

Try stringr:

library(stringr)
str_extract_all(c('abc','1b2b3b4'), '.b.')
# [[1]]
# [1] "abc"
# 
# [[2]]
# [1] "1b2" "3b4"

I can't believe nobody ever mentioned regmatches!

x <- c('abc','1b2b3b4')
regmatches(x, gregexpr('.b.', x))

# [[1]]
# [1] "abc"

# [[2]]
# [1] "1b2" "3b4"

It makes me wonder, didn't regmatches exist two and half years ago?

You should probably give Gabor Grothendieck the check for writing the gsubfn package:

 require(gsubfn)
#Loading required package: gsubfn
 strapply(c('abc','1b2b3b4'), ".b.", I)

#Loading required package: tcltk
#Loading Tcl/Tk interface ... done
[[1]]
[1] "abc"

[[2]]
[1] "1b2" "3b4"

This just applies the identity function , I, to the matches of the pattern.

You need to combine gregexpr with substring, I reckon:

> s = c('abc','1b2b3b4')
> m = gregexpr('.b.',s)
> substring(s[1],m[[1]],m[[1]]+attr(m[[1]],'match.length')-1)
[1] "abc"
> substring(s[2],m[[2]],m[[2]]+attr(m[[2]],'match.length')-1)
[1] "1b2" "3b4"

The returned list 'm' has the start and lengths of matches. Loop over s to get all the substrings.