#!/usr/bin/perl
use strict;
use warnings;

my $foo;   # Replace with $_ accordingly

my %data;
my %table = ( 94 => "A8A40" );

my @fields = qw(Year Month Day Request_Seq Request_Type Record_Seq Rec
+ord_Data);

@data{ @fields } = unpack "A2A2A2A4A2A2A1486" , $foo;
my ($Destination, $DestinationType) = unpack $table{$data{Record_Type}
+} , $data{Record_Data};
[download]

Beauty!

That is exactly the direction that I needed. My code is now like this:

my %Field;
my @HeaderFields=qw(RequestYear RequestMonth RequestDay RequestSequenc
+e RecordType RecordSequence Record);
my %RecordMap =(Preface => A2A2A2A4A2A2A1486,
               94      => A8A40);
...
@Field{@HeaderFields}=unpack("$RecordMap{Preface}",$_);
...
[download]

Now just to change the rest of the mess into the nice clean hashes. (but first, lunch!)

Thanks LR (and all the other monks)

Skip

In order to avoid potential mismatches of the field name and pack template values, I would think about using a hash to generate both the unpack template and the field names. The only problem is hashes are not stored in any particular order. To combat this, one could store the "hash" in an array, and separate the keys and values "manually." Example:

my @fields = qw(Year         A2
                Month        A2 
                Day          A2
                Request_Seq  A4
                Request_Type A2
                Record_Seq   A2
                Record_Data  A1486);

my @keys   = map $fields[$_], grep $_%2==0, 0 .. $#fields;
my @values = map $fields[$_], grep $_%2==1, 0 .. $#fields;

my %data; # bad name. we should know better. ;)
@data{ @keys } = unpack join("", @values), $foo;
[download]

Alternatives to this include, but are not limited to:

• Use a tied hash that maintains order, such as Tie::IxHash.
• Use a normal hash, but also keep along an array that is used to sort the hash keys and values. This is essentially what a tied hash would do for you, automatically.
• Some other extremely clever solution I have not thought of.

my %unpack = ( 
preface => "A2 A2 A2 A4 A2 A2 A1486",
94 => "A8 A40",
...
)
@HeaderRecord   = unpack $unpack{preface}, $FullRecord;

@RecordNames[$RecordType] =  unpack $unpack{$RecordType}, $Record;
[download]

Ah, but I really want the names. Keeping track of exactly what field 35 for record type 22 is what I want to avoid. I think tracking that I am working on a NameAddress record and reading HomeZipCode is going to be a bit easier to maintain.

Thanks,

Skip

@RecordNames[94]       = qw/$Destination  $DestinationType/;
@RecordLengths[94]     = qw/$A8            A4/;
[download]

$RecordNames[94] = "$Destination";
$RecordLengths[94] = "$A8";
[download]

$fields[94]  = [ qw/Destination DestinationType/ ];
$lengths[94] =   qq/A8          A4             /  ;

my @data{ @{$fields[94]} } = unpack($lengths[94], $Record);

# then you have:
#
# $data{Destination}
# $data{DestinationType}
[download]

my $assign;
$assign="(".join(",",@HeaderRecordNames).")=";
$assign.="unpack(\".join("",@RecordLengths[$RecordType])."\");";
print "Evaluating $assign\n"; # this is for debugging
eval $assign;
[download]

I would like to split the packing/unpacking function from the data definition so that I can put it all in one place, and reuse it.

Use a subroutine.

For what it's worth, here's what I usually do when confronted with stuff like this: (well, sort of - I actually have a few modules now that make this kind of thing easier)

my %recordformat =
(
  '__COMMON__'  => [qw(
                       RequestYear    1-2
                       RequestMonth   3-4
                       RequestDay     5-6
                       RequestSequence 7-10
                       RecordType     11-12
                       RecordSequence 13-14
                     )],
   94           => [qw(
                       Destination     15-22
                       DestinationType 23-26
                     )]
   # other stuff
);

sub fillHashFromData
{
  my ($destination, $recordSpec, $data) = @_;
  my @spec = @$recordSpec; # make a copy
  while (@spec)
  {
    my $key = shift @spec;
    my $location = shift @spec;
    $location =~ /(\d+)(?:-(\d))?/ or
        die "Syntax error in recordSpec: @$recordSpec";
    my ($start, $end) = ($1, $2 || $1);
    $destination->{$key} = substr($data, $start - 1, $end - $start + 1
+);
  }
}

# Much later....

my $dataLine;
while ($dataLine = <VENDORFILE>)
{
  my %datavalues;
  fillHashFromData(\%datavalues, $recordformat{__COMMON__}, $dataLine)
+;
  fillHashFromData(\%datavalues, $recordformat{$datavalues{'RecordType
+'}}, $dataLine);

# Do stuff here

}
[download]

  {
    no strict qw(vars refs);
    while(my ($k,$v) = each %datavalues) {$$k = $v;}
  }
[download]

One distinct advantage of using a %recordformat hash in the format above is that a section of it can be cut-and-pasted directly from the vendor's documentation; also, it becomes easy for you to remove a huge number of fields that you're ignoring from consideration without making certain that you have the right number of "x" specifications in the pack string. (Occasionally vendors will give field lengths; in my experience, they always give character position ranges)

The subroutine that lets you construct a data line from a hash and a format is also pretty easy:

sub fillDataFromHash
{
  my ($src, $recordSpec, $data) = @_;
  $data ||= ' ' x $RECORD_LENGTH;  # Maybe add to %recordformat
  my @spec = @$recordSpec;
  while (@spec)
  {
    my $key = shift @spec;
    my $location = shift @spec;
    $location =~ /(\d+)(?:-(\d))?/ or
        die "Syntax error in recordSpec: @$recordSpec";
    my ($start, $end) = ($1, $2 || $1);
    substr($data, $start - 1, $end - $start + 1) = $src->{$key};
  }
}

# ...
# fill %datavalues somehow
my $dataLine = fillDataFromHash(\%datavalues, 
                                $recordformat{__COMMON__});
$dataLine = fillDataFromHash(\%datavalues, 
                             $recordformat{$datavalues{'RecordType'}},
+ $dataLine);
[download]

@HeaderRecordNames   = qw/$RequestYear $RequestMonth $RequestDay $Requ
+estSequence $RecordType $RecordSequence $Record/;
@HeaderRecordLengths = qw/A2           A2            A2          A4   
+            A2          A2              A1486/;

...

my (@HeaderRecordNames) = unpack(@HeaderRecordLengths, $_);
my (@RecordNames[$Recordtype])=unpack(@RecordLengths[$RecordType]);
[download]

@HeaderRecordNames   = qw/RequestYear RequestMonth RequestDay RequestS
+equence RecordType RecordSequence Record/;
$HeaderRecordLengths = q/A2A2A2A4A2A2A1486/;

...

my @head = unpack($HeaderRecordLengths, $_);
my %Header = map {$HeaderRecordNames[$_] => $head[$_]} (0..$#head);
my @data = unpack($RecordLengths[$head{RecordType}], $Header{Record});
my %Record = map {$RecordNames[$_] => $data[$_]} (0..$#data);
[download]