help with regex

mmittiga17 has asked for the wisdom of the Perl Monks concerning the following question:

Hi All,

need to pull just date out of lines. Date is preceded by (PURCHASED ON|SOLD ON|SETTLED ON). example CLASS A SOLD ON 17/11/10 FOR SETTLEMENT ON 22/11/10 @ 145.59993 just want to grab 17/11/10 as $tdate and 22/11/10 as $sdate

while(defined($line = (<IN>))){
          if ($line =~m/(PURCHASED ON|SOLD ON|SETTLEMENT ON)/) {
              $desc = (split(/\,/, $line))[27]; 
              #print "$desc\n";
              $tdate = (($desc =~ /PURCHASED\s+ON\s+(.*?)\s+/g )||($de
+sc =~ /SOLD\s+ON\s+(.*?)\s+/g ));
              $sdate = $desc =~m/SETTLEMENT\S\s+ON\s+(.*?)\s+/g ;
             print "TDATE $tdate SDATE $sdate\n";
          }      
   }
[download]

Any help will be appreciated.

Comment on help with regex Download Code

Replies are listed 'Best First'.
Re: help with regex by AnomalousMonk (Archbishop) on Nov 30, 2010 at 00:56 UTC
Another way (note variables `$tdate $sdate` are undefined, hence easy to test, if no match exists): `>perl -wMstrict -le "my $line = 'CLASS A SOLD ON 17/11/10 FOR SETTLEMENT ON 22/11/10 @ 145.59993'; ;; my $date = qr{ \d\d / \d\d / \d\d }xms; ;; my ($tdate) = $line =~ m{ (?: PURCHASED \| SOLD) \s+ ON \s+ ($date) }xmsg; print qq{tdate: '$tdate'}; ;; my ($sdate) = $line =~ m{ SETTLEMENT \s+ ON \s+ ($date) }xmsg; print qq{sdate: '$sdate'}; " tdate: '17/11/10' sdate: '22/11/10'` [download] See also perlretut, perlrequick, perlreref. Update: Or even (hash `%dates` empty if no match): `>perl -wMstrict -le "use Data::Dumper; ;; my $date = qr{ \d\d / \d\d / \d\d }xms; my $sep = qr{ \s+ ON \s+ }xms; my @ops = qw(PURCHASED SOLD SETTLEMENT); my $op = join '\|', map quotemeta, @ops; $op = qr{ $op }xms; ;; READ_LOOP: { ;; my $line = 'CLASS A SOLD ON 17/11/10 FOR SETTLEMENT ON 22/11/10 @ 145.59993'; ;; next READ_LOOP if not my %dates = $line =~ m{ ($op) $sep ($date) }xmsg; print Dumper \%dates; ;; } " $VAR1 = { 'SETTLEMENT' => '22/11/10', 'SOLD' => '17/11/10' };` [download]	[reply] [d/l] [select]
Re: help with regex by toolic (Bishop) on Nov 30, 2010 at 00:36 UTC
`while (<DATA>) { if (/(PURCHASED ON\|SOLD ON\|SETTLEMENT ON)/) { $desc = $_; if ($desc =~ /SETTLEMENT\s+ON\s+(\S+)/) { $sdate = $1; } if ($desc =~ /(?:PURCHASED\|SOLD)\s+ON\s+(\S+)/) { $tdate = $1; } print "TDATE $tdate SDATE $sdate\n"; } } __DATA__ CLASS A SOLD ON 17/11/10 FOR SETTLEMENT ON 22/11/10 @ 145.59993` [download] Prints: `TDATE 17/11/10 SDATE 22/11/10` [download] perlre	[reply] [d/l] [select]