HL7 v2 Parser

Decode a pipe-delimited HL7 v2.x message segment by segment, 100 % client-side. Type, version, encoding characters, field-by-field table.

How it works

HL7 v2 (v2.5.1 and earlier) serialises messages as segments separated by carriage returns. Each segment opens with a 3-character identifier (MSH, PID, EVN…) followed by |-separated fields.

The parser runs in two phases:

1. Read the MSH segment: the 4^th character defines the field separator (MSH-1); the next 4 characters define the component / repetition / escape / sub-component separators (MSH-2). All 5 must be unique.
2. Tokenise the remaining segments using those characters. Every field is broken into repetitions, components and sub-components so positional access such as PID-5.1.1 remains stable.

Encoding characters

The five HL7 v2.x encoding characters defined in chapter 2 section 2.5 must all be unique:

• | — field separator (MSH-1).
• ^ — component separator.
• ~ — repetition separator.
• \ — escape character (for \F\, \S\, \R\, \E\, \T\).
• & — sub-component separator.

A message that deviates from those defaults (for example $%&\#) remains valid as long as the 5 characters stay unique. The parser reads whatever values appear in MSH-1 and MSH-2.

Recognised message types

The parser does not enforce a type: it extracts the message type and trigger event from MSH-9 (code^triggerEvent^messageStructure) and the version from MSH-12. The most common types are:

• ADT — Admit / Discharge / Transfer (admissions, transfers, discharges).
• ORM / OMG / OMD / OMP — Orders (lab, imaging, pharmacy).
• ORU — Observations / results.
• SIU — Scheduling Information Unsolicited (appointments).
• BAR — Add / Change Billing Account.
• MFN — Master File Notification.

Limitations

• The parser does not validate the content of non-MSH segments. PID-3, for instance, may contain anything from this tool's standpoint.
• Escape sequences (\F\, \S\, …) are preserved verbatim — you see the escaped character exactly as sent.
• The tool does not strip the MLLP framing (preamble \x0B / trailer \x1C\x0D): paste only the textual payload between the MLLP boundaries.
• Segment groupings (e.g. PID/{ NK1 } inside ADT_A01) are not reconstructed — the segment list is flat.

Privacy

Related documentation

• HL7 v2.x — the standard, its history and ecosystem.
• Official spec: HL7 v2.5.1 product brief.
• FHIR R5 — the RESTful successor to HL7 v2, in JSON / XML.