Access regex group results by name using gregexpr in R

Question

How can I access all matches with groupings (i.e. label) using gregexpr()in R?

s <- "aaa123bbb345ccc"
p <- "(?<label>\\d+)"
m <- gregexpr(p, s, perl = TRUE)

I am interested in printing the matches in m but referencing the group name <label1>. This can be easily done in C# but I'm struggling in R and I cannot figure out how to do this from the CRAN docs.

Edit: C# code below requested by G Grothendieck:

string s = "aaa123bbb345ccc";
string p = @"(?<label>\d+)";
Regex r = new Regex(p);
Match m = r.Match(s);
if (m.Success)
{
    Console.WriteLine(m.Groups["label"].Value);
}

["The best way to use regular expressions with R is to pass the perl=TRUE parameter. This tells R to use the PCRE regular expressions library."](http://www.regular-expressions.info/rlanguage.html) — marekful, Dec 02 '16 at 22:34
`regmatches(s, m)` or avoid the awkward two-function approach with `stringr::str_extract_all(s, p)` — alistaire, Dec 02 '16 at 22:37
When you use `gregexpr` with `regmatches`, only matches are preserved. The easiest is to use stringr `str_match_all`. `regexec` could be an alternative, but it does not work with PCRE regexps. — Wiktor Stribiżew, Dec 02 '16 at 22:40
Maybe a dupe of [Regex group capture in R with multiple capture-groups](http://stackoverflow.com/questions/952275/regex-group-capture-in-r-with-multiple-capture-groups) — Wiktor Stribiżew, Dec 02 '16 at 22:53

G. Grothendieck · Accepted Answer · 2016-12-02T23:20:55.203

1

This will return only the substrings associated with label.

st <- attr(m[[1]], "capture.start")[, "label"]
len <- attr(m[[1]], "capture.length")[, "label"]
substring(s, st, st + len - 1)
## [1] "123" "345"

edited Dec 02 '16 at 23:20

answered Dec 02 '16 at 22:55

G. Grothendieck

254,981
17
203
341

Thank you, this is close. I'm getting an "Error in capture.start[, "label"] : subscript out of bounds". – Hahnemann Dec 02 '16 at 23:17
I have revised the answer a few times. Check the latest one. It works for me if run after the code supplied in the question. – G. Grothendieck Dec 02 '16 at 23:19
It works! This is what I was looking for. I wish this would be easier in R like it is in C# but it provides what I need. – Hahnemann Dec 02 '16 at 23:27
Can you add some code at the bottom of your question providing an example of C# code that does it. – G. Grothendieck Dec 02 '16 at 23:28

Access regex group results by name using gregexpr in R

1 Answers1