提取数字之间的单词

Question

Trying to write some regex in R to extract some words between numbers for each string in a character vector in R. Unfortunately, my regex skills aren't nearly up to the challenge.
Here's an example of the problem and my initial attempt:

x <- c("1 Singleword 1,234 342", "2 randword & thirdword 1,545 323", 
      "3 Anotherword wordagain Newword. 3,234 556")

m <- regexpr("[a-zA-Z]+\\s+", x, perl = TRUE)

regmatches(x, m)

This approach only produces

"Singleword ", "randword ", "Anotherword "

What I need is

"Singleword", "randword & thirdword", "Anotherword wordagain Neword."

I believe it will need to be some kind of regex pattern that will start with a character (like I currently have) and then pull everything until a number is reached.

Answer 1

x <- c("1 Singleword 1,234 342", "2 randword & thirdword 1,545 323", 
       "3 Anotherword wordagain Newword. 3,234 556")

m <- regexpr("[a-zA-Z].(\\D)+", x, perl = TRUE)

regmatches(x, m)

[1] "Singleword " "randword & thirdword "
[3] "Anotherword wordagain Newword. "

I used https://regexr.com/ and it's cheatsheet to figure out how to compose the regex.

Answer 2

Using sub

> sub(".\\s(\\D+).*", "\\1", x)
[1] "Singleword "   "randword & thirdword "  "Anotherword wordagain Newword. "

Using str_extract

> library(stringr)
> str_extract(x, pattern = "\\D+")
[1] " Singleword "  " randword & thirdword "  " Anotherword wordagain Newword. "

Answer 3

sample data

x <- c("1 Singleword 1,234 342", "2 randword & thirdword 1,545 323", 
   "3 Anotherword wordagain Newword. 3,234 556")

Base R

#replace als numbers and comma's with `""` (=nothing), 
# also, trim whitespaces (thanks Markus!)
trimws( gsub( "[0-9,]", "", x ) )

[1] "Singleword" "randword & thirdword" "Anotherword wordagain Newword."

stringR

library(stringr)
str_extract(x, pattern = "(?<=\\d )[^0-9]+(?= \\d)")

[1] "Singleword" "randword & thirdword" "Anotherword wordagain Newword."

If you like to learn more about (the working of) regex-patterns in the code above (and in the other answers), check out their magic (and explanation) at: https://regex101.com/

explanation of the last regex-pattern: https://regex101.com/r/QgERuZ/2