cochrasc has asked for the wisdom of the Perl Monks concerning the following question:

Hey all. I stink at regexes, so I thought I would grab some help. I would normally use a split to do something like this, but the data isn't always consistent, so I think a regex is my only option. If there are any other options, I'd be happy to hear them.
Here's the dealy....I'm trying to break down a string that can contain several (unknown) instances of 5 pieces of data.
Here's an example of one that contains 3 instances:
value:patternList = "{error 1 1 {^E [0-2][0-9]:[0-5][0-9]:[0-5][0-9].*
+$} {^E [0-2][0-9]:[0-5][0-9]:[0-5][0-9].*$}} {three 1 1 {^.*35=A.*$|^
+.*35=5.*$} {^.*35=A.*$|^.*35=5.*$}} {fixv 1 1 ^.*VFIXFxProxy.*Disconn
+ected ^.*VFIXFxProxy.*Disconnected}"

[download]
The first instance starts with "error", the 2nd "three" and the 3rd "fixv". Notice that the fixv instance doesn't have "{}" around the 4th and 5th pieces of data like the other 2 do...this is my problem and why I can't use split. Also note that the lack of "{}" can occur in any of the instances, not just the 3rd.
So...in a nutshell...I'm trying to pull out the 3 instances into an array which I can then split (or use another regex) to get my 5 pieces of data from each.

The data itself is separated by spaces (with data containing a space encompassed with {}), so in instance 1, the 5 pieces would be: 
error
1
1
^E [0-2][0-9]:[0-5][0-9]:[0-5][0-9].*$
^E [0-2][0-9]:[0-5][0-9]:[0-5][0-9].*$

[download]
I hope this makes sense...it's difficult to explain. Thanks in advance for any help.

CODE tags added by Arunbear

Replies are listed 'Best First'.
Re: Need Regex help
by Roy Johnson (Monsignor) on Sep 09, 2005 at 14:41 UTC
    You should use <code></code> tags around your code, so your posts don't get mangled.

    It looks like you want to pull out outer-level matched braces. You might want to look at Text::Balanced.

    split takes a regex, by the way.

    Caution: Contents may have been coded under pressure.
[reply]
[d/l]
Re: Need Regex help
by JediWizard (Deacon) on Sep 09, 2005 at 14:36 UTC

    Try this:

    #!/usr/local/bin/perl

$re = qr@
             \{
             (?:
                (?> [^{}]+ )    # Non-{} without backtracking
              |
                (??{ $re })     # Group with matching {}
             )*
             \}
          @x;

my $string = 'value:patternList = "{error 1 1 {^E 0-20-9:0-50-9:0-50-9
+.*$} {^E 0-20-9:0-50-9:0-50-9.*$}} {three 1 1 {^.*35=A.*$|^.*35=5.*$}
+ {^.*35=A.*$|^.*35=5.*$}} {fixv 1 1 ^.*VFIXFxProxy.*Disconnected ^.*V
+FIXFxProxy.*Disconnected}"';

while($string =~ m/$re/g){
    print "$&\n";
}

exit;

__END__

output:
{error 1 1 {^E 0-20-9:0-50-9:0-50-9.*$} {^E 0-20-9:0-50-9:0-50-9.*$}}
{three 1 1 {^.*35=A.*$|^.*35=5.*$} {^.*35=A.*$|^.*35=5.*$}}
{fixv 1 1 ^.*VFIXFxProxy.*Disconnected ^.*VFIXFxProxy.*Disconnected}

    [download]

    See perlre

    They say that time changes things, but you actually have to change them yourself.

    —Andy Warhol

[reply]
[d/l]
      So using this, how would he then get his five elements from the results? Seeing as how some of the elements are encased in {} if they contain spaces or |'s? 
      {error 1 1 {^E 0-20-9:0-50-9:0-50-9.*$} {^E 0-20-9:0-50-9:0-50-9.*$}}
{three 1 1 {^.*35=A.*$|^.*35=5.*$} {^.*35=A.*$|^.*35=5.*$}}
{fixv 1 1 ^.*VFIXFxProxy.*Disconnected ^.*VFIXFxProxy.*Disconnected}

      [download]
[reply]
[d/l]

        #!/usr/local/bin/perl

$re = qr@
             \{(
             (?:
                (?> [^{}]+ )    # Non-{} without backtracking
              |
                (??{ $re })     # Group with matching {}
             )*
             )\}
          @x;

my $string = 'value:patternList = "{error 1 1 {^E 0-20-9:0-50-9:0-50-9
+.*$} {^E 0-20-9:0-50-9:0-50-9.*$}} {three 1 1 {^.*35=A.*$|^.*35=5.*$}
+ {^.*35=A.*$|^.*35=5.*$}} {fixv 1 1 ^.*VFIXFxProxy.*Disconnected ^.*V
+FIXFxProxy.*Disconnected}"';

my $count = 1;
while($string =~ m/$re/g){
        my $inst = $1;
        my(@elements) = ($inst =~ m/((?<={)[^}]+(?=})|[^\s{}]+)/g);
        print "Instance $count\'s elements = ".join("\n", @elements)."
+\n\n\n";
        $count++;
}

exit;


__END__
Instance 1's elements = error
1
1
^E 0-20-9:0-50-9:0-50-9.*$
^E 0-20-9:0-50-9:0-50-9.*$


Instance 2's elements = three
1
1
^.*35=A.*$|^.*35=5.*$
^.*35=A.*$|^.*35=5.*$


Instance 3's elements = fixv
1
1
^.*VFIXFxProxy.*Disconnected
^.*VFIXFxProxy.*Disconnected

        [download]

        Update Removed unnessicary grouping paranethesis.

        They say that time changes things, but you actually have to change them yourself.

        —Andy Warhol

[reply]
[d/l]
      this regexp doesn't seem to be working as posted.
[reply]

        What version of perl are you using? It works as posted for me running perl 5.6.1 on Redhat linux 8.

        From perlre:

        (??{ code })
    WARNING: This extended regular expression feature is considered hi
+ghly experimental, and may be changed or deleted without notice. A si
+mplified version of the syntax may be introduced for commonly used id
+ioms. 

    This is a ``postponed'' regular subexpression. The code is evaluat
+ed at run time, at the moment this subexpression may match. The resul
+t of evaluation is considered as a regular expression and matched as 
+if it were inserted instead of this construct.

    The code is not interpolated. As before, the rules to determine wh
+ere the code ends are currently somewhat convoluted.

    The following pattern matches a parenthesized group:

  $re = qr{
             \(
             (?:
                (?> [^()]+ )    # Non-parens without backtracking
              |
                (??{ $re })     # Group with matching parens
             )*
             \)
          }x;

        [download]

        They say that time changes things, but you actually have to change them yourself.

        —Andy Warhol

[reply]
[d/l]
Re: Need Regex help
by philcrow (Priest) on Sep 09, 2005 at 14:50 UTC
    I would start with the extract_bracketed function from Text::Balanced to grab the bits in the braces. Then you can treat the pieces you get from that individually.

    Phil

[reply]

Back to Seekers of Perl Wisdom