first two words-pattern matching

[shamala]

Hi monks, i have a file whose contents are as below:

<first line> :ctx     ctx_3   ctx_3   040425          0       0       
+07-May-2004 03:29      07-May-2004 03:29        0.00    hobbit       
+   N/A     failed          PLE-QA N/A      0
<second line>        Not Published   466008691
<third line>plsql   tkmain_ps4      tkmain_ps4      040425          0 
+      0       07-May-2004 03:29       07-May-2004 03:29       0.00   
+ hobbit          N/A     failed PLE-QA   N/A     0
<...>        Not Published   466008691
<nth line>plsql   tkmain_pu3      tkmain_pu3      040425          0   
+    0       07-May-2004 03:29       07-May-2004 03:29       0.00    h
+obbit          N/A     failed PLE-QA   N/A     0
        Not Published   466008691
rdbms   tkmain_2n1      tkmain_2n1      040425          0       0     
+  07-May-2004 01:37       07-May-2004 01:37       0.00    hobbit     
+     N/A     failed PLE-QA   N/A     0
        Not Published   465894765

I want only the first two words(for ex:ctx ctx_53) from each of the lines that begin with the similar kind of words like the third line has - plsql tkmain_ps4 and i want this to be matched. I have written the following but isnt working...could any of you please help...i am a begginner at perl.

#! /usr/local/bin/perl
      #use strict;
       open(rerun,"rerunlrg")
       while(my $pattern=<rerun>){
        if(my $pattern=~ m/^(\w\s+\w\s).*?/){
        print "matched\n";
        print "$1\n";
        }

Thanks ppl

Edited by Chady -- added code tags around file contents.

[davido]

Can we assume that <nth line> really is just your notation for the beginning of a new line, where lines end with "\n"?

If that's the case:

use strict;    # Don't comment it out.
open RERUN, "rerunlrg" or die "Couldn't open input file.\n$!";
while ( my $line = <RERUN> ) {
    print "Matched:\n$1\n\n" if $line =~ m/^(\S+\s+\S+)/;
}
close RERUN;

Go ahead and give that a try.

One problem your regexp has is that you're not putting a quantifier after your \w metachars. That means that you're trying to match a single word character, followed by any positive amount of whitespace, followed by a single word character, followed by a single whitespace, followed by any amount of anything. You really want to be matching words of arbitrary length, presumably longer than a single character each.

Another problem your script has is that you're not checking for success when you open the file. You should probably be invoking die if there is a failure to open the file. See the example I provided. You can read up on this issue in perlopentut.

[matija]

^(\w\s+\w\s) That matches a single character folowed by one or more spaces, folowed by another single character, folowed by one space.

What you probably want is something along the lines of

^(\w+\s+\w+\b)

Note that I removed the final space from your match by replacing it with \b - I assume you don't really need that final space.

[TilRMan]

use strict;
use warnings;

open RERUN, "rerunlrg" or die "Can't open rerunlrg: $!";
while (<RERUN>) {
  next if $. % 2 == 0; # Skip the second, fourth, ... line

  if (m[(\S+)\s+(\S+)]) {
    print "Matched: $1 $2\n";
  }
}
close RERUN;

[shamala]

Hey TilRman ..dat was neat!!! thanks

[saskaqueer]

while (<DATA>) {

    # We could use split( /\s+/, $_, 3 ) to explicitly
    # set the max split limit to 3, but split() is smart
    # enough to see we are capturing into 2 variables and
    # automatically sets the limit to 3. Smart perl!

    my ($first, $second) = split( /\s+/ );
    print "Line $.: '$first $second'\n";
}

__DATA__
this is the first line right here
line two comes next way down here
then third line is here
fourth comes next

[snadra]

Hello,

your file seems to be a csv file. Wich is not seperated by commas but by tabs. But the seperator is not important. Maybe the Text::CSV module or other modules about csv can help you as well.

snadra