Top level class for generating u.n. edi interchange objects and subobjects.
use Business::EDI; my $edi = Business::EDI-new('d09b'); # set the EDI spec version my $rtc = $edi->codelist('ResponseTypeCode', $json) or die "Unrecognized code!"; printf "EDI response type: %s - %s (%s)\n", $rtc->code, $rtc->label, $rtc->value; my $msg = Business::EDI::Message->new($ordrsp) or die "Failed Message constructor"; foreach ($msg->xpath('line_detail/all_LIN') { ($_->part(7143) || '') eq 'EN' or next; print $_->part(7140)->value, "\n"; # print all the 13-digit (EN) ISBNs }
The focus of functionality is to provide object based access to \s-1EDI\s0 messages and subelements. At present, the \s-1EDI\s0 input processed by Business::EDI objects is \s-1JSON\s0 from the edi4r ruby library, and there is no \s-1EDI\s0 output beyond the perl objects themselves.
When you \*(C`use Business::EDI;\*(C' the following package namespaces are also loaded:
Business::EDI::Segment_group Business::EDI::Message
That's why the example message constructor in \s-1SYNOPSIS\s0 would succeed without having done \*(C`use Business::EDI::Message;\*(C'
Everything depends on the spec. That means you have to have declared a spec version before you can create or parse a given chunk of data. The exception is a whole \s-1EDI\s0 message, because each message declares its spec version internally.
\s-1EDI\s0 has a hierachical specification defining data. From top to bottom, it includes:
This module handles messages and everything below, but not (yet) communications.
Much more documentation needed here...
Constructor
Get/set accessor for the value of the field.
The string code designating this node's type. The code is what is what the spec uses to refer to the object's definition. For example, a composite \*(L"C504\*(R", segment \*(L"\s-1RFF\s0\*(R", data element \*(L"7140\*(R", etc.
Don't be confused when dealing with CodeList objects. Calling code() gets you the 4-character code of the CodeList field, \s-1NOT\s0 what that CodeList is currently set to. For that use value().
English description of the element.
This method returns strings that can be fed to part() like:
foreach ($x->part_keys) { something($x->part($_)) }
This is similar to doing:
foreach (keys %x) { something($x{$_}) }
In this way an object can be exhaustively, recursively parsed without further knowledge of it.
Returns subelement(s) of the object. The key can reference any subobject allowed by the spec. If the subobject is repeatable, then prepending \*(L"all_\*(R" to the key will return an array of all such subobjects. This is the safest and most comprehensive approach. Using part($key) without \*(L"all_\*(R" to retrieve when there is only one $key subobject will succeed. Using part($key) without \*(L"all_\*(R" to retrieve when there are multiple $key subobjects will \s-1FAIL\s0. Since that difference is only dependent on data, you should always use \*(L"all_\*(R" when dealing with a repeatable field (or xpath, see below).
Examples:
my $qty = $detail->part('QTY'); # FAILURE PRONE! my @qtys = $detail->part('all_QTY'); # OK!
$path can traverse multiple depths in representation via one call. For example:
$message->xpath('all_SG26/all_QTY/6063')
is like this function foo():
sub foo { my @x; for my $sg ($message->part->('all_SG26') { for ($sg->part('all_QTY') { push @x, $->part('6063'); } } return @x; }
The xpath version is much nicer! However this is nowhere near as fully featured as W3C xpath for \s-1XML\s0. This is more like a multple-depth part().
Examples:
my @obj_1154 = $message->xpath('line_detail/SG31/RFF/C506/1154');
Returns value(s) instead of object(s).
Examples:
'\s-1ORDRSP\s0' eq $ordrsp->xpath_value('\s-1UNH/S009/0065\s0') or die \*(L"Wrong Message Type!\*(R";
This code is experimental. \s-1EDI\s0 is a big spec with many revisions.
At the lower levels, all data elements, codelists, composites and segments from the most recent spec (D09B) are present.
Business::EDI::Spec edi4r - http://edi4r.rubyforge.org
Joe Atzberger