EDIFACT field access

Read and write EDIFACT business data through semantic Python names instead of positions.

Typed EDIFACT messages expose their segments, elements and components as Python attributes with semantic names - the same model as HL7 v2 field access. The names come from the dialect's message classes, so every dialect you define per Dialects and profiles reads the same way, with the names you choose.

Segments and elements

Each segment a message declares is an attribute of the message, and each element is an attribute of its segment:

from zato.edifact import parse_edifact

# The Dutch healthcare dialect, shipped as an example
import zato.edifact.nl

msg = parse_edifact(raw).message

# A segment, then an element on it
sex = msg.pid.sex

# Elements that hold a composite have typed components
family_name = msg.pid.patient_name.married_name
prefix      = msg.pid.patient_name.married_name_prefix
born_year   = msg.pid.date_of_birth.year

A message class maps pid to the wire tag PID and patient_name to the element position where the name composite lives, so services read like the business logic they implement.

Composites

Composite data elements group their components under one attribute:

# An address composite and its components
street   = msg.pad.address.street
number   = msg.pad.address.building_number
city     = msg.pad.address.city
postcode = msg.pad.address.postcode

Repeating segments

A segment declared as repeatable is a list - iterate it, index it, take its length:

# TXT repeats - one segment per line of text
for line in msg.text:
    print(line.text)

first_line = msg.text[0].text
line_count = len(msg.text)

Repeating groups

Groups model the nested, repeating structures of larger messages. In a lab result, for example, each specimen has its own determinations, and each determination its own remarks:

for material in msg.materials:

    print(material.det.date.day, material.ide.identification_number)

    for determination in material.determinations:
        print(determination.determination, determination.result, determination.unit)

    for remark in material.determination_remarks:
        print(remark.text)

The service segments

UNH and UNT are typed like any other segment, and the interchange envelope segments UNB and UNZ are reachable on the interchange itself:

interchange = parse_edifact(raw)
msg = interchange.message

message_type = msg.unh.identifier.message_type
version      = msg.unh.identifier.version
reference    = msg.unh.reference_number
count        = msg.unt.segment_count

sender    = interchange.header.sender.identification
recipient = interchange.header.recipient.identification

Positional fallback

You can read any segment - typed or not - by position, which covers fields outside the declared structure and wholly unknown segments:

# All segments with a given tag, in wire order
for seg in msg.segments('NAD'):

    # e_1, e_2 .. resolve to elements by position
    print(seg.e_1)

Building and modifying

Segments are plain classes - instantiate one, assign its elements and serialize it:

from zato.edifact.nl.segments import TXT

line = TXT()
line.text = 'Please review the attached results'

wire = line.serialize()   # TXT+Please review the attached results'

Assignments on a parsed message merge into the original wire data during serialization - what you change is re-serialized, everything else stays as received.

Page	What it covers
Message parsing	How the typed objects these names live on come into being
Dialects and profiles	Where the semantic names are defined and how to choose your own
EDIFACT in healthcare tutorial	Typed access in a complete service, from lab result to reply