regexpr and only matching prices and not other digits

Question

I'm trying to come up with code that will extract only the price from a line of text.

Motivated by RegEx for Prices?, I came up with the following command:

gregexpr('\\d+(\\.\\d{1,2})', '23434 34.232 asdf 3.12  ')

[[1]]
[1]  7 19
attr(,"match.length")
[1] 5 4
attr(,"useBytes")
[1] TRUE

However, in my case, I would only like 3.12 to match and not 34.232. Any suggestions?

If you want to extract the value, then its better/straightforward to use `sub` than `gregexpr`. — Arun, Feb 18 '13 at 07:43

Michael · Accepted Answer · 2013-02-18T03:26:43.460

3

I think this should work:

'\\d+\\.\\d{1,2}(?!\\d)'

edited Feb 18 '13 at 03:26

answered Feb 18 '13 at 03:20

Michael

Thanks, works for me when used with `perl = T` argument. Above also works with `gregexpr`. – andrewj Feb 19 '13 at 03:33

Explosion Pills · Answer 2 · 2013-02-19T11:39:08.543

2

\\d+\\.\\d{1,2}(?!\\d)

I'm not 100% sure that negative lookahead is supported in r, so here is an alternative:

\\d+\\.\\d{1,2}(?:[^\\d]|$)

edited Feb 19 '13 at 11:39

answered Feb 18 '13 at 03:22

Explosion Pills

3

Things like negative lookaheads are supported in R if you use the `perl=TRUE` argument. – Marius Feb 18 '13 at 03:24
1

Wouldn't you need to wrap `\\d+\\.\\d{1,2}` in parenthesis if you wanted to be able to capture it? – Michael Feb 18 '13 at 03:35
1

@Michael Not if it's the entire match. – Explosion Pills Feb 18 '13 at 03:37
Hi, I just tried the alternative, and it's matching `34.232`, not `3.12`. – andrewj Feb 19 '13 at 03:36
@andrewj Michael was correct about capturing, since the final digit will also be captured by the whole match – Explosion Pills Feb 19 '13 at 11:39

AndrewP · Answer 3 · 2013-02-18T04:47:32.753

-2

one or more digits followed by a point, followed by 1 or 2 digits, followed by white space or end of string

\\d+\\.\\d{1,2}(\w|$)

Edit: as per comments, R uses double-escape

edited Feb 18 '13 at 04:47

answered Feb 18 '13 at 03:20

AndrewP

Won't work. You didn't escape the period. You're also either capturing words after the end of the price or assuming that the price is at the end of the string, both of which the OP probably doesn't want. . – Michael Feb 18 '13 at 03:22
Right you are about the escaping of period. accidentally un-escaped it... wrapping the entire thing in parenthesis and putting it as part of a bigger statement will allow Marius to just select the matching token – AndrewP Feb 18 '13 at 03:28
1

Note the `r` tag. That's why all the other answers have double escapes: \\ . – Matthew Lundberg Feb 18 '13 at 03:41
@Matthew this is the difficulty in tagging regex within R. regex in a way is its own thing and those unfamiliar with R assume regex is regex. I'm often hesitant about using the regex tag in R (on SO) for this reason. – Tyler Rinker Feb 18 '13 at 04:31
Noted, thanks for alerting me to paying more attention to Tag descriptions :) – AndrewP Feb 18 '13 at 04:46

3 Answers3