Re: [Boost-users] Newbie conversion of C# to boost::regex

From: Andrew Holden (aholden_at_[hidden])
Date: 2007-03-12 11:57:16

Jonathon Douglas wrote:
>"John Maddock" <john_at_[hidden]> wrote in message 
>news:00b701c7648e$1b5e3180$69d36b51_at_fuji...
>> Jonathon Douglas wrote:
>>> I am trying to convert the following C# code over to using
>>> boost::regex
>>>
>>> MatchCollection col = new
>>> Regex(@"\w+").Matches(@"slx://account/123456"); string table =
>>> col[1].Value;
>>> string id = col[2].Value;
>>>
>>> Table would equal "account" and id would equal "123456" in this.
>>>
>>> Ive tried using regex_match with this but it always finds no
matches.
>>>
>>> Any sugestions ?
>>
>> C++ gobbles up \'s in starings, so it looks you want "\\w+" so that
the
>> regex engine gets to see \w+
>>
>> John.
>
>
>Hi,
>
>Ive tried that, however been a complete newbie to boost im probably
still 
>missing something obvious. Heres the test code so far
>
>#include <stdlib.h>
>#include <iostream>
>#include <string>
>#include <boost/regex.hpp>
>using namespace std;
>int main (int argc, char *argv[])
>{
>string uri ("slx://account/12345");
>boost::regex re ("\\w+");
>boost::cmatch what;
>if (boost::regex_match(uri.c_str(), what, re))
>{
>cout << what[0] << endl;
>cout << what[1] << endl;
>cout << what[2] << endl;
>}
>else
>{
>cout << "nowt matched" << endl;
>}
>return 0;
>} 

It looks like there is some confusion about boost::regex_match does.  It
attempts to match the entire string against the regular expression
(imagine an implicit ^ and $ around the expression).

I am not familiar with the C# library, but it looks like "Matches" finds
all of the matches in the string and returns an array of them.  Is this
correct?

If it is, then there are two things you can do:

Option 1: Change your expression to use captures.  It might look like
"slx://(\\w+)/(\\w+)".  If you drop this ion your code above, you should
see that what[0] is the entire uri, what[1] is "account" and what[2] is
"12345".  I generally prefer this method, as it also provides syntax
checking.

Option 2: Use boost::regex_iterator.  Perhaps something like:

//WARNING:  completely untested code.  I am just threw this together
from information I found in
http://www.boost.org/libs/regex/doc/regex_iterator.html

string uri ("slx://account/12345");
boost::regex re ("\\w+");

boost::sregex_iterator ri (uri.begin(), uri.end(), re);
boost::sregex_iterator end;   //An uninitialized regex_iterator works as
end

while (ri != end)
{
        //Dereference the iterator to get a boost::match_results object,
then take match_results[0] to get the whole match.
        std::cout << (*ri)[0] << std::endl;

        ++ri;  //Advance to the next match
}