From: Robert Klemme <bob.news@...>
Date: 2005-12-20T06:47:50+09:00
Subject: Re: regular expressions question

ako... <akonsu@gmail.com> wrote:
> yes, thank you. this is a better description of the problem. i am not
> a native english speaker, so may be this is one of the reasons why my
> question is not clear.
>
> i saw a solution to this problem that uses split at the end. it of
> course won't work if you change your example and allow quoted strings
> in source-words and destination-words. a quoted string can contain
> anything, spaces too and your keywords too, so the subsequent split
> won't work.
>
> well, i did not realise that the term "group's captures" is that rare.
> i thought it was a standard term. but may be i am brainwashed by
> microsoft. so i have this code in .net which might help to clarify
> what i am talking about:
>
>            string text = "One car red car blue car";
>            string pat = @"^(?:(\w+)\s+)*(\w+)$";
>            Regex r = new Regex(pat, RegexOptions.IgnoreCase);
>
>            // Match the regular expression pattern against a text
> string.
>            Match m = r.Match(text);
>            if (m.Success)
>            {
>                Console.WriteLine("match: [{0}]", m);
>                foreach (Group g in m.Groups)
>                {
>                    Console.WriteLine("group: [{0}]", g);
>                    foreach (Capture c in g.Captures)
>                    {
>                        Console.WriteLine("\tcapture: [{0}]", c);
>                    }
>                }
>            }
>
> the output is:
>
> match: [One car red car blue car]
> group: [One car red car blue car]
>        capture: [One car red car blue car]
> group: [blue]
>        capture: [One]
>        capture: [car]
>        capture: [red]
>        capture: [car]
>        capture: [blue]
> group: [car]
>        capture: [car]
>
> as you see, the first group is $0, the second group is $1, and the
> third is $2. but $1 and $2 contain captures too. it is like if $1 and
> $2 were arrays in Ruby.
>
> in my opinion this is a big limitation of ruby's regular expressions.
> it just must be as powerful as .net ; -)
>
> konstantin

I don't know whether your question was answered in the lengthy thread 
already.  In case not: in Ruby to get all matches of a group you need to 
iterate through the whole string with #scan.  There is no such thing as this 
feature of .net - and frankly I haven't missed it so far.  To get at all the 
words in your example this is sufficient:

>> s = "One car red car blue car"
=> "One car red car blue car"
>> s.scan /\w+/
=> ["One", "car", "red", "car", "blue", "car"]

If you actually need group matches, you'll have to do something like this

>> s.scan(/\w(\w+)/).map{|m| m[0]}
=> ["ne", "ar", "ed", "ar", "lue", "ar"]

alternative

>> ma=[]
=> []
>> s.scan(/\w(\w+)/) {|m| ma << m[0]}
=> "One car red car blue car"
>> ma
=> ["ne", "ar", "ed", "ar", "lue", "ar"]

Of course this is quite a silly example...  The main point here is that you 
must refrain from anchoring the regexp at the beginning if you want to 
iterate like this.

HTH

    robert