Well, there are two really. First, every XML file will have the same general structure but not the same information. For example, the XML file A may have a social security number while B doesn't have one. Or, study C may contain EKG readings that don't show up in A or B. Using DOM, I would have to build a level of intelligence and redundancy to ensure that the script doesn't crap out if it doesn't see certain information. From what I've read, the event driven model is more appropriate, since I getting data fed from a another server.

the other is speed. On my local workstation, the script runs fine. But this module will be part of a (slightly) larger automated script for a hospital in which multiple patients (XML files) will input at the same time. Like most patient care systems, I can't afford for this thing to crash b/c of memory hogging.

Here is the code:

sub CamPatData
{
    my $obj = shift;
    Date_Init("TZ=EST");

    
    if (@_)
    {
        my $file = shift;
        my $parser = XML::DOM::Parser->new();
        my $doc = $parser->parsefile($file);
        my (@patdemo);
        
        my ($ln, $fn, $sex, $dob, $htu, $htn, $wtu, $wtn, $patnum, $pa
+tient);
        
        #Get Patient Data
        foreach $patient($doc->getElementsByTagName('SA_PATIENT'))
        {
                    
            $ln = $patient->getElementsByTagName('MY_LAST_NAME')->item
+(0)->getFirstChild->getNodeValue;
            $fn = $patient->getElementsByTagName('MY_FIRST_NAME')->ite
+m(0)->getFirstChild->getNodeValue; 
            $dob = $patient->getElementsByTagName('MY_BIRTH_DATE')->it
+em(0)->getFirstChild->getNodeValue;
            $sex = $patient->getElementsByTagName('MY_GENDER')->item(0
+)->getFirstChild->getNodeValue; 
            $dob = $patient->getElementsByTagName('MY_BIRTH_DATE')->it
+em(0)->getFirstChild->getNodeValue;    
        }
            
            
            
            
        foreach $patient($doc->getElementsByTagName('CathStudy'))
        {
                    
            $htn = $patient->getElementsByTagName('PatHeight')->item(0
+)->getFirstChild->getNodeValue;
            $htu = $patient->getElementsByTagName('PatHeight')->item(0
+)->getAttribute('Units');
            $wtn = $patient->getElementsByTagName('PatWeight')->item(0
+)->getFirstChild->getNodeValue; 
            $wtu = $patient->getElementsByTagName('PatWeight')->item(0
+)->getAttribute('Units');
            $patnum = $patient->getElementsByTagName('StudyID')->item(
+0)->getFirstChild->getNodeValue;
            
        }        
        
        #Rounding Off 

        $wtn = ceil($wtn);
        $htn = ceil($htn);
        
        my $date = ParseDate($dob);    
        $dob = UnixDate($date,"%m/%d/%Y");
        
        @patdemo = ($ln, $fn, $dob, $sex, $htn, $htu, $wtn, $wtu, $pat
+num);

}
[download]

Here is the Sample Data:

</CathStudy>

<SA_PATIENT>
<MY_PATIENT_ID>XXXXX</MY_PATIENT_ID>
<MY_BIRTH_DATE>123-45-6789</MY_BIRTH_DATE>

<MY_FIRST_NAME>BERT</MY_FIRST_NAME>

<MY_GENDER>M</MY_GENDER>

<MY_LAST_NAME>PUPPET</MY_LAST_NAME>

<Address>123 SESAME ST</Address>

<City>NEW YORK</City>

<State>NY</State>

<Zip>10110</Zip>

<Race>Caucasian</Race>

</SA_PATIENT>
[download]

Any suggestions?

Looks like the simplest way to do this might be to slurp all the data into a hash using XML::Simple then either clean up the hash or extract the data from it into another hash as is appropriate:

use strict;
use warnings;
use Data::Dump::Streamer;
use XML::Simple;

my $str = do {local $/; <DATA>};
my %data = %{XMLin ($str)};

Dump (\%data);
[download]

Read more... Data and output (2 kB)