Regex: Match first group of whitespaces

Archdove · 2012-06-13 09:57:21

Hi

I'm trying to match the first group of whitespaces in a string in order to split the string there.
What i try to accomplish is separating the arguments from a command:

  mono   program.exe arg1 arg2
  mono___program.exe arg1 arg2

My regex matches the first group of whitespaces after the first word and all the following groups:

(?<=\w)\s+
    mono___program.exe__arg1__arg2

I just cant figure out to prevent the matches between ...exe___arg1___arg2.

The match should work with any number of arguments 0 - n and also with other programs next to mono.

Any idea? You can work you regex here:
http://gskinner.com/RegExr/

Last edited by Archdove (2012-06-13 10:02:08)

/dev/zero · 2012-06-13 10:33:26

The question is ill-posed. I shouldn't need to go to a website to test a regex. I already have grep, awk and lex installed, just like every other Unix machine on the planet - why can't I use them?

Archdove · 2012-06-13 11:16:16

I did never say you cant use grep, awk or whatever. I simply said that you can work the regex there.
For people like me, with no mad regex skills, it's quite convenient to see the matching graphically. (yes vim ... yes)

But over all, this ensures a common base for all of the people helping so we (i) don't face problems with differing regex standards.

As to my ill-posted question ... Really?

Question:
How can i match the the first group of white spaces after the first word? Please refer to post 1.

theDOC · 2012-06-13 11:36:03

How about:

^[^ ]+([ ]+)(.*)

?

Archdove · 2012-06-13 11:44:24

@theDOC: This seems to match everything or nothing if there are preceding white-spaces.

theDOC · 2012-06-13 11:54:49

In your case it matches

mono
<spaces>
program.exe arg1 arg2

isn't that correct?

Archdove · 2012-06-13 12:07:54

Well. What you're saying is correct yes. But the match seems to be:

nothing on " mono program.exe foo bar "
everything on "mono program.exe foo bar"

I tried it with the above posted website. Which tool do you use to to debug regex?

theDOC · 2012-06-13 12:22:08

Well, I don't know the range of your possible input lines...
This will skip (possible) preceding spaces:

^[ ]*[^ ]+([ ]+)(.*)

PS: I tried with your posted website, too.

fracek · 2012-06-13 13:16:00

Do you have to use RegEx or you can use something else? I wrote a simple (and vastly improvable) state machine that does what I thing you want to achieve.

Example Gist (in javascript)

#!/usr/bin/env node

/*
 * Given a string return a map in the form { cmd: 'args', cmd2: 'args' }
 */
function parseCommand(s) {
  m = {}
  state = 0;
  currentcmd = '';
  currentargs = '';
  for (i = 0; i < s.length; i++) {
    c = s[i];
    //console.log(c, ' ', state);
    switch (state) {
      case 0:
        if (c != ' ') {
          state = 1;
          currentcmd += c;
        }
        break;
      case 1:
        if (c == ' ') {
          state = 2;
        } else {
          currentcmd += c;
        }
        break;
      case 2:
        if (c != ' ') {
          state = 3;
          currentargs += c;
        }
        break;
      case 3:
        if (c == '\n') {
          state = 0;
          m[currentcmd] = currentargs;
          currentcmd = '';
          currentargs = '';
        } else {
          currentargs += c;
        }
        break;
    }
  }
  m[currentcmd] = currentargs;
  return m;
}

var cmds = ['mono program.exe arg1 arg2',
            '   mono program.exe arg1 arg2',
            'cmd1 arg1 arg2 arg3\n    cmd2        arg11 arg12 etc    '];
cmds.forEach(function(c) {
  console.log(c, parseCommand(c));
});

alphaniner · 2012-06-13 14:28:25

Regular expressions stop at the first match unless told to do otherwise, so all you should need is \s+

Edit: If that doesn't suit you, maybe (?<=(^\w*))\s+ would work. It depends if lookarounds (thanks for introducing me to this btw) can be variable length in whatever tool you're using (can't in grep ).
Edit: Also can't in perl... so much for that.

But I don't understand what you hope to accomplish by matching the whitespace. Isn't the stuff before and after the whitespace what you really want?

Last edited by alphaniner (2012-06-13 14:58:14)

Archdove · 2012-06-14 16:44:02

I did it manually like facek suggested but much simpler. I just tought that doing it with regex would be fun but it isn't that easy it seems.

Why i tried to match the whitespaces is that i want to split the string at that position leaving mit with an array of two strings. (with a single instruction)
Matching either part of the string just felt like going the long way around since two instructions would be necessary.

So thanks for all the help. Regex is just th wrong tool for accomplishing what i wanted.

debdj · 2012-06-15 11:28:57

I couldn't really understand what you want to achieve.

but from what I understand, if you want mono , program.exe arg1 arg2 from 'mono program.exe arg1 arg2'
then I'll do something like

a='mono program.exe arg1 arg2'
args=${a#* } #$args will contain 'program.exe arg1 arg2'
prog=${a%% *} #$prog will contain 'mono'

you can find some explanations here: http://www.ibm.com/developerworks/linux … index.html

But if that is at all what you want..

Trilby · 2012-06-15 11:34:59

or to get arg1 and arg2

awk '// {print $3 $4;}'

But then this is sounding more and more like a homework question by the minute.

Arch Linux

#1 2012-06-13 09:57:21

Regex: Match first group of whitespaces

#2 2012-06-13 10:33:26

Re: Regex: Match first group of whitespaces

#3 2012-06-13 11:16:16

Re: Regex: Match first group of whitespaces

#4 2012-06-13 11:36:03

Re: Regex: Match first group of whitespaces

#5 2012-06-13 11:44:24

Re: Regex: Match first group of whitespaces

#6 2012-06-13 11:54:49

Re: Regex: Match first group of whitespaces

#7 2012-06-13 12:07:54

Re: Regex: Match first group of whitespaces

#8 2012-06-13 12:22:08

Re: Regex: Match first group of whitespaces

#9 2012-06-13 13:16:00

Re: Regex: Match first group of whitespaces

#10 2012-06-13 14:28:25

Re: Regex: Match first group of whitespaces

#11 2012-06-14 16:44:02

Re: Regex: Match first group of whitespaces

#12 2012-06-15 11:28:57

Re: Regex: Match first group of whitespaces

#13 2012-06-15 11:34:59

Re: Regex: Match first group of whitespaces

Board footer