Common mistakes in EDI files

The most frequent technical reasons behind rejected messages, and how to spot them.

When an EDI message is rejected, the cause usually lies in syntax or qualifier choice rather than business logic. These are the ones seen most often in practice.

1. The release character is forgotten

A +, : or ' appearing inside a free-text field (FTX, IMD) is taken for a separator and splits the line. The fix is to precede such characters with ?.


FTX+AAI+++DELIVERY AFTER 08?:00      ← correct
FTX+AAI+++DELIVERY AFTER 08:00       ← breaks the line

2. Empty elements are dropped

In EDIFACT, meaning comes from position. An unused element cannot be removed; it is left empty in place:


NAD+SU+8712345678901::9++ACME FOODS BV    ← correct (the double + preserves it)
NAD+SU+8712345678901::9+ACME FOODS BV     ← the name lands in the wrong element

3. The UNT segment count does not match

UNT carries the number of segments in the message, including UNH and UNT themselves. Adding a line by hand without updating the count makes the receiving system reject the message as a checksum failure.

4. Wrong date format code

The third element of DTM states the format. Confusing 102 with 203 makes the time part read as if it belonged to the date:


DTM+137:20260117:102'      ← 17 Jan 2026
DTM+137:202601171030:203'  ← 17 Jan 2026 10:30

5. Missing currency

PRI and MOA carry no currency; it comes from the CUX segment at message level. Without a CUX the amounts are ambiguous and the receiver applies its own default — on invoices, a mistake with a real cost.

6. MOA qualifiers get mixed up

The summary section holds several MOA segments, each a different total: 79 goods total, 124 tax, 77 grand total, 9 amount payable. Treating the first MOA as the total without reading the qualifier is a common error.

7. Character set mismatch

UNOA in the UNB accepts only uppercase letters and digits. Sending text with accented characters as UNOA corrupts it at the other end; UNOC should be used for Latin-1.

How to check

Opening the file in ediviewer highlights the segment tags and lists every element of the selected line with a name against it. An element sitting opposite a label you did not expect usually points to one of the mistakes above.