Extract numbers between brackets within a string

Question

Possible Duplicate:
Extract info inside all parenthesis in R (regex)

I inported data from excel and one cell consists of these long strings that contain number and letters, is there a way to extract only the numbers from that string and store it in a new variable? Unfortunately, some of the entries have two sets of brackets and I would only want the second one? Could I use grep for that?

the strings look more or less like this, the length of the strings vary however:

"East Kootenay C (5901035) RDA 01011"

or like this:

"Thompson-Nicola J (Copper Desert Country) (5933039) RDA 02020"

All I want from this is 5901035 and 5933039

Any hints and help would be greatly appreciated.

Is it possible for there to be two instances of numbers-in-brackets in the same line? For example, `"East Kootenay C (5901035) (5933039) RDA 01011"` — Blue Magister, Oct 04 '12 at 20:52

Justin · Answer 1 · 2012-10-04T22:23:38.843

There are many possible regular expressions to do this. Here is one:

x=c("East Kootenay C (5901035) RDA 01011","Thompson-Nicola J (Copper Desert Country) (5933039) RDA 02020")

> gsub('.+\\(([0-9]+)\\).+?$', '\\1', x)
[1] "5901035" "5933039"

Lets break down the syntax of that first expression '.+\$([0-9]+)\$.+'

.+ one or more of anything
\\( parentheses are special characters in a regular expression, so if I want to represent the actual thing ( I need to escape it with a \. I have to escape it again for R (hence the two \s).
([0-9]+) I mentioned special characters, here I use two. the first is the parentheses which indicate a group I want to keep. The second [ and ] surround groups of things. see ?regex for more information.
?$ The final piece assures that I am grabbing the LAST set of numbers in parens as noted in the comments.

I could also use * instead of . which would mean 0 or more rather than one or more i in case your paren string comes at the beginning or end of a string.

The second piece of the gsub is what I am replacing the first portion with. I used: \\1. This says use group 1 (the stuff inside the ( ) from above. I need to escape it twice again, once for the regex and once for R.

Clear as mud to be sure! Enjoy your data munging project!

I'd add two characters to your regex to make sure that OP is grabbing the last number-in-bracket: `gsub('.+\$([0-9]+)\$.+?$', '\\1', x)`. Perhaps swap out `+` and put in `*` depending on needs as well. — Blue Magister, Oct 04 '12 at 20:54

score 3 · Answer 2 · answered Oct 04 '12 at 23:00

Here is a gsubfn solution:

library(gsubfn)

strapplyc(x, "[(](\\d+)[)]", simplify = TRUE)

[(] matches an open paren, (\\d+) matches a string of digits creating a back-reference owing to the parens around it and finally [)] matches a close paren. The back-reference is returned.

Extract numbers between brackets within a string

2 Answers2

Linked

Related