[R] regexpr help (match.length=0)

[Erik Iverson]

McGehee, Robert wrote:
> R-help,
> Sorry if this is more of a regex question than an R question. However,
> help would be appreciated on my use of the regexpr function.
> 
> In the first example below, I ask for all characters (a-z) in 'abc123';
> regexpr returns a 3-character match beginning at the first character. 
> 
>> regexpr("[[:alpha:]]*", "abc123")
> [1] 1
> attr(,"match.length")
> [1] 3
> 
> However, when the text is flipped regexpr, and I ask for a match of all
> characters in '123abc', regexpr returns a zero-character match beginning
> at the first character. Can someone explain what a zero length match
> means (i.e. why not return -1), and why the result isn't 4,
> match.length=3?

It means it matches 0 characters, which is fine since you use *, which 
means match 0 or more occurrences of the regex.  It sounds like you want 
+ instead of *.  Also see gregexpr.
> 
>> regexpr("[[:alpha:]]*", "123abc")
> [1] 1
> attr(,"match.length")
> [1] 0
>