GESMES — EDIFACT Generic Statistical Message (D.02B)

The GESMES message is the generic carrier of statistical series between statistical offices, central banks and international bodies (ECB, Eurostat, BIS, IMF): it transports structured time-series with their dimensions, codes and values, without presupposing any business domain — a cross-domain vehicle that preceded SDMX. This page summarises the segment structure as published by UN/CEFACT for directory D.02B.

Purpose

The GESMES message is the generic carrier of statistical series between statistical offices, central banks and international bodies (ECB, Eurostat, BIS, IMF): it transports structured time-series with their dimensions, codes and values, without presupposing any business domain — a cross-domain vehicle that preceded SDMX.

This page focuses on the structure published by UN/CEFACT for directory D.02B. The segments listed, their status (M / C), position and repetition factor are extracted from the official file gesmes_c.htm hosted on service.unece.org.

Segment structure

The GESMES D.02B message has 88 header entries, 0 detail entries and 0 summary entries (segments and groups combined). Groups are shown in bold and their members are indented by actual nesting depth.

Header

Detail

No segment in this section.

Summary

No segment in this section.

Changes vs D.96A

Top-level differences between D.96A and D.02B for the GESMES message (segments directly inside header / detail / summary, ignoring nested groups). The comparison is indicative: a given tag may have evolved inside nested groups without showing up in this synthesis.

Segments added vs D.96A

• GEI in the header section (status C, repeat ×5) — not present at the top of the corresponding D.96A section.

Segments removed vs D.96A

• GIS in the header section — present in D.96A but missing in D.02B at the same level.

Max-repeat changes

• ARR (header): repeat ×9999 in D.96A → ×999999 in D.02B.

Structural example

Minimal example of an GESMES D.02B message with mandatory segments only. The goal is to show the exact version declared in UNH (GESMES:D:02B:UN) — not to replicate a full business case (see the GESMES D.96A page for an annotated example).

UNB+UNOC:3+5410000000123:14+5410000000456:14+260514:1430+CTRL002001'
UNH+1+GESMES:D:02B:UN'
BGM+14+GESMES-2026-Q1-FR+9'
DTM+137:20260514:102'
DTM+9:20260331:102'
BUS++17'
RFF+SRS:EUROSTAT-DS-2026-CPI'
UNT+7+1'
UNZ+1+CTRL002001'

Common errors

• Misaligned UNH tokens — The UNH+...+GESMES:D:02B:UN token must reflect exactly the directory in use — a mismatch between the UNH release and the release agreed with the partner is rejected by every strict validator.
• Wrong UNT segment count — The value after UNT+ must include UNH and UNT themselves. By far the most common sender-side error, regardless of the directory in use.
• Stale code lists — Between D.96A and D.02B, several code lists received new qualifiers (DTM status, RFF qualifier, NAD codes…). Reusing an older directory’s code tables triggers warnings or even rejections depending on the partner’s strict-conformance mode.

Related messages

Other versions of this message available on ediverse.io: GESMES D.96A, GESMES D.01B, GESMES D.10A, GESMES D.16B, GESMES D.21B, GESMES D.24A, GESMES D.97A, GESMES D.99B, GESMES D.00B, GESMES D.05A, GESMES D.08A, GESMES D.13B, GESMES D.18B, GESMES D.20B.

← All EDIFACT versions