Matching over two lines

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

I have a feeling that I'm overlooking something obvious, but I've been away from Perl for about a year, and I'm having trouble getting the code below to work.

I'm trying to read through a system log file, and pull out the users that have entered orders on a patient. The way to tell that, however, is a two-line process...

I need to find the line that has a segment beginning with 'UB' (user identifier) directly following a pipe... but only when it's following a line (and not necessarily the next line) that has 'MSG-HDR' in it.

Here's what I've got so far:

$orderfound = 0;     # Flag used to keep track of Orders being found

while (<>) {
  $line = $_;

  if ($line=~/MSG-HDR/)
     { $orderfound = 1; 
     }
      
  if (($orderfound) && ($line=~/|UB/))
     { print "$line \n";
       $orderfound = 0; 
     } 
}
[download]

And here's a file snippet... (names changed, obviously)

 *-X1AA1 ENTERED* C 6 18 374,032906  P 090003 090003  23-027   D,0    
+                  043800 VORHEES, JASON           
 *-X1AA1 ENTERED* C 6 18 374,032906  P 090003 090003  43-044   D,0    
+                  045800 MYERS, MICHAEL           
 *-X1AA1 ENTERED* C 6 18 373,032906  P 090003 090003  21-032   D,0    
+                  047000 KREUGER, FREDDY           
0* * * * * * * * * * * * * * * * * * * * * * * * * *  36-060 090003   
+                                                   
 *MSG-HDR  D8020DE9 00000056 143CA421 007C0000 0329A008 84090003      
+                                                <01
 |U E9C160  ||UBZ 0000  VORHEES, JASON              RPH  SJP ||UD  000
+0. M8||QA# 0000  ||QA= 11FF  ||QB= 1130  Q0V|    <02
 |QB= 00C8  QA3|                                                      
+                                                <03
 * * * * * * * * * * * * * * * * * * * * * * * * * *                  
+                                                   
 *-Q1AA1 ENTERED* C 6 18 371,032906  P 090003 090003  36-060   D,0    
+                  046800 VORHEES, JASON
[download]

What's hapening right now is that I'm printing out the line that has 'MSG-HDR' in it, but I need it to print the line with 'UBZ 0000 VORHEES, JASON' in it.

I'm sure I'm missing something obvious... I just don't see it...

Comment on Matching over two lines Select or Download Code

Replies are listed 'Best First'.
Re: Matching over two lines by ikegami (Patriarch) on Mar 30, 2006 at 17:00 UTC
Three problems: `"\|"` is a special character in regexps. It needs to be escaped. You will look for `"MSG-HDR"` and `"\|UB"` on the same line. The following would match: `MSG-HDR \|UB` [download] The `next` in the code below fixes this. `$orderfound` is not always cleared when it should be. The following would match: `MSG-HDR bla \|UB` [download] Or maybe that's ok? If it's ok, keep the `$orderfound = 0;` where it was. Fix: `my $orderfound = 0; while (my $line = <>) { if ($line =~ /MSG-HDR/) { $orderfound = 1; next; } if ($orderfound && $line =~ /\\|UB/) { print "$line\n"; } $orderfound = 0; }` [download]	[reply] [d/l] [select]
Re^2: Matching over two lines by TrekNoid (Pilgrim) on Mar 30, 2006 at 17:34 UTC
Escaping the pipe did the trick The reason I'm clearing $orderfound where I am is because the UB line could be 1 or more lines later than the MSG-HDR line, so I don't want $orderfound to clear until I've matched that line. Thanks much...	[reply]